Merge pull request #1423 from MODSetter/dev

feat: improved agent speed and fixed it citations

.github/workflows/backend-tests.yml

@@ -4,6 +4,9 @@ on:
   pull_request:
     branches: [main, dev]
     types: [opened, synchronize, reopened, ready_for_review]
+    paths:
+      - 'surfsense_backend/**'
+      - '.github/workflows/backend-tests.yml'
 
 concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
@@ -21,26 +24,15 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v6
 
-      - name: Check if backend files changed
-        id: backend-changes
-        uses: dorny/paths-filter@v3
-        with:
-          filters: |
-            backend:
-              - 'surfsense_backend/**'
-
       - name: Set up Python
-        if: steps.backend-changes.outputs.backend == 'true'
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.12'
 
       - name: Install UV
-        if: steps.backend-changes.outputs.backend == 'true'
-        uses: astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@v8.1.0
 
       - name: Cache dependencies
-        if: steps.backend-changes.outputs.backend == 'true'
         uses: actions/cache@v5
         with:
           path: |
@@ -51,19 +43,16 @@ jobs:
             python-deps-
 
       - name: Cache HuggingFace models
-        if: steps.backend-changes.outputs.backend == 'true'
         uses: actions/cache@v5
         with:
           path: ~/.cache/huggingface
           key: hf-models-${{ env.EMBEDDING_MODEL }}
 
       - name: Install dependencies
-        if: steps.backend-changes.outputs.backend == 'true'
         working-directory: surfsense_backend
         run: uv sync
 
       - name: Run unit tests
-        if: steps.backend-changes.outputs.backend == 'true'
         working-directory: surfsense_backend
         run: uv run pytest -m unit
 
@@ -93,26 +82,15 @@ jobs:
       - name: Checkout code
         uses: actions/checkout@v6
 
-      - name: Check if backend files changed
-        id: backend-changes
-        uses: dorny/paths-filter@v3
-        with:
-          filters: |
-            backend:
-              - 'surfsense_backend/**'
-
       - name: Set up Python
-        if: steps.backend-changes.outputs.backend == 'true'
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.12'
 
       - name: Install UV
-        if: steps.backend-changes.outputs.backend == 'true'
-        uses: astral-sh/setup-uv@v7
+        uses: astral-sh/setup-uv@v8.1.0
 
       - name: Cache dependencies
-        if: steps.backend-changes.outputs.backend == 'true'
         uses: actions/cache@v5
         with:
           path: |
@@ -123,19 +101,16 @@ jobs:
             python-deps-
 
       - name: Cache HuggingFace models
-        if: steps.backend-changes.outputs.backend == 'true'
         uses: actions/cache@v5
         with:
           path: ~/.cache/huggingface
           key: hf-models-${{ env.EMBEDDING_MODEL }}
 
       - name: Install dependencies
-        if: steps.backend-changes.outputs.backend == 'true'
         working-directory: surfsense_backend
         run: uv sync
 
       - name: Run integration tests
-        if: steps.backend-changes.outputs.backend == 'true'
         working-directory: surfsense_backend
         env:
           TEST_DATABASE_URL: postgresql+asyncpg://postgres:postgres@localhost:5432/surfsense_test

.github/workflows/code-quality.yml

@@ -11,13 +11,13 @@ concurrency:
 
 jobs:
   file-quality:
-    name: File Quality Checks
+    name: File Quality
     runs-on: ubuntu-latest
     if: github.event.pull_request.draft == false
     
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v6
         with:
           fetch-depth: 0
 
@@ -27,7 +27,7 @@ jobs:
           git fetch origin ${{ github.base_ref }}:${{ github.base_ref }} 2>/dev/null || git fetch origin ${{ github.base_ref }} 2>/dev/null || true
 
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.12'
 
@@ -35,7 +35,7 @@ jobs:
         run: pip install pre-commit
 
       - name: Cache pre-commit hooks
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pre-commit
           key: pre-commit-${{ hashFiles('.pre-commit-config.yaml') }}
@@ -74,7 +74,7 @@ jobs:
     
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v6
         with:
           fetch-depth: 0
 
@@ -83,7 +83,7 @@ jobs:
           git fetch origin ${{ github.base_ref }}:${{ github.base_ref }} 2>/dev/null || git fetch origin ${{ github.base_ref }} 2>/dev/null || true
 
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.12'
 
@@ -91,7 +91,7 @@ jobs:
         run: pip install pre-commit
 
       - name: Cache pre-commit hooks
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pre-commit
           key: pre-commit-security-${{ hashFiles('.pre-commit-config.yaml') }}
@@ -125,35 +125,36 @@ jobs:
           exit ${exit_code:-0}
 
   python-backend:
-    name: Python Backend Quality
+    name: Backend Quality
     runs-on: ubuntu-latest
     if: github.event.pull_request.draft == false
     
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v6
         with:
           fetch-depth: 0
 
       - name: Set up Python
-        uses: actions/setup-python@v5
+        uses: actions/setup-python@v6
         with:
           python-version: '3.12'
 
       - name: Install UV
-        uses: astral-sh/setup-uv@v3
+        uses: astral-sh/setup-uv@v8.1.0
 
       - name: Check if backend files changed
         id: backend-changes
-        uses: dorny/paths-filter@v3
+        uses: dorny/paths-filter@v4
         with:
           filters: |
             backend:
               - 'surfsense_backend/**'
+              - '.github/workflows/code-quality.yml'
 
       - name: Cache dependencies
         if: steps.backend-changes.outputs.backend == 'true'
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: |
             ~/.cache/uv
@@ -171,7 +172,7 @@ jobs:
 
       - name: Cache pre-commit hooks
         if: steps.backend-changes.outputs.backend == 'true'
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pre-commit
           key: pre-commit-backend-${{ hashFiles('.pre-commit-config.yaml') }}
@@ -206,13 +207,13 @@ jobs:
           exit ${exit_code:-0}
 
   typescript-frontend:
-    name: TypeScript/JavaScript Quality
+    name: Frontend Quality
     runs-on: ubuntu-latest
     if: github.event.pull_request.draft == false
     
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v6
         with:
           fetch-depth: 0
 
@@ -221,24 +222,24 @@ jobs:
           git fetch origin ${{ github.base_ref }}:${{ github.base_ref }} 2>/dev/null || git fetch origin ${{ github.base_ref }} 2>/dev/null || true
 
       - name: Setup Node.js
-        uses: actions/setup-node@v4
+        uses: actions/setup-node@v6
         with:
-          node-version: '18'
+          node-version: '20'
 
       - name: Install pnpm
-        uses: pnpm/action-setup@v4
-        with:
-          version: latest
+        uses: pnpm/action-setup@v6
 
       - name: Check if frontend files changed
         id: frontend-changes
-        uses: dorny/paths-filter@v3
+        uses: dorny/paths-filter@v4
         with:
           filters: |
             web:
               - 'surfsense_web/**'
+              - '.github/workflows/code-quality.yml'
             extension:
               - 'surfsense_browser_extension/**'
+              - '.github/workflows/code-quality.yml'
 
       - name: Install dependencies for web
         if: steps.frontend-changes.outputs.web == 'true'
@@ -254,7 +255,7 @@ jobs:
         run: pip install pre-commit
 
       - name: Cache pre-commit hooks
-        uses: actions/cache@v4
+        uses: actions/cache@v5
         with:
           path: ~/.cache/pre-commit
           key: pre-commit-frontend-${{ hashFiles('.pre-commit-config.yaml') }}

.pre-commit-config.yaml

@@ -67,7 +67,7 @@ repos:
       # Biome check for surfsense_web
       - id: biome-check-web
         name: biome-check-web
-        entry: bash -c 'cd surfsense_web && npx @biomejs/biome check --diagnostic-level=error .'
+        entry: bash -c 'cd surfsense_web && npx @biomejs/biome@2.4.6 check --diagnostic-level=error .'
         language: system
         files: ^surfsense_web/
         pass_filenames: false

VERSION

@@ -1 +1 @@
-0.0.24
+0.0.25

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/citations/on.md

@@ -1,11 +1,42 @@
 <citations>
-Apply chunk citations only when the runtime injects `<document>` /
-`<chunk id='…'>` blocks.
+Citations reach the answer through two channels. Use whichever applies — and
+never invent ids you didn't see. Citation ids are resolved by exact-match
+lookup; a wrong id silently breaks the link, so when in doubt, omit.
+
+### Channel A — chunk blocks injected this turn
+When `search_surfsense_docs` or `web_search` returns `<document>` /
+`<chunk id='…'>` blocks in this turn:
 
 1. For each factual statement taken from those chunks, add
-   `[citation:chunk_id]` using the exact id from `<chunk id='…'>`.
-2. Multiple chunks → `[citation:id1], [citation:id2]` (comma-separated).
-3. Never invent or normalise ids; if unsure, omit.
-4. Plain brackets only — no markdown links, no footnote numbering.
-5. If no chunk-tagged documents appear this turn, do not fabricate citations.
+   `[citation:chunk_id]` using the **exact** id from a visible
+   `<chunk id='…'>` tag. Copy digit-for-digit (or the URL verbatim);
+   do not retype from memory.
+2. `<document_id>` is the parent doc id, **not** a citation source —
+   only ids inside `<chunk id='…'>` count.
+3. Multiple chunks → `[citation:id1], [citation:id2]` (comma-separated,
+   each id copied individually).
+4. Never invent, normalise, or guess at adjacent ids; if unsure, omit.
+5. Plain brackets only — no markdown links, no footnote numbering.
+
+### Channel B — citations relayed by a `task` specialist
+A `task(...)` tool message may contain `[citation:<chunk_id>]` markers
+the specialist already attached to its prose. The specialist saw the
+underlying `<chunk id='…'>` blocks; you didn't. So:
+
+1. **Preserve those markers verbatim** in your final answer — do not
+   reformat, renumber, drop, or wrap them in markdown links. When you
+   paraphrase a specialist sentence, copy the marker character-for-
+   character; do not regenerate the id from memory (LLMs reliably
+   corrupt nearby digits).
+2. Keep each marker attached to the sentence the specialist attached
+   it to.
+3. Do **not** add new `[citation:…]` markers of your own to a
+   specialist's prose; if a fact has no marker, the specialist
+   couldn't tie it to a chunk and neither can you.
+4. When a specialist returns JSON, the citation markers live inside
+   the prose-bearing fields (e.g. a summary or excerpt). Pull them
+   along with the surrounding sentence when you quote.
+
+If neither channel surfaces citation markers this turn, do not fabricate
+them.
 </citations>

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/memory_protocol/private.md

@@ -6,4 +6,10 @@ standing instructions?
 If yes, call `update_memory` **alongside** your normal response — don't
 defer it to a later turn. Skip ephemeral chat noise (one-off Q/A, greetings,
 session logistics). Stay within the budget shown in `<user_memory>`.
+
+Memory is heading-based markdown. New entries should be under `##` headings
+such as `## Facts`, `## Preferences`, or `## Instructions`, with bullets like
+`- YYYY-MM-DD: text`. If existing memory contains legacy
+`(YYYY-MM-DD) [fact|pref|instr]` markers, preserve the information but write
+new saves in the heading-based format.
 </memory_protocol>

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/memory_protocol/team.md

@@ -6,4 +6,12 @@ key facts?
 If yes, call `update_memory` **alongside** your normal response — don't
 defer it to a later turn. Skip ephemeral chat noise (one-off Q/A, greetings,
 session logistics). Stay within the budget shown in `<team_memory>`.
+
+Team memory is heading-based markdown. New entries should be under `##`
+headings such as `## Product Decisions`, `## Engineering Conventions`,
+`## Project Facts`, or `## Open Questions`, with bullets like
+`- YYYY-MM-DD: text`. If existing memory contains legacy `(YYYY-MM-DD) [fact]`
+markers, preserve the information but write new saves in the heading-based
+format. Do not create personal headings such as `## Preferences` or
+`## Instructions`.
 </memory_protocol>

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/tools/update_memory/private/description.md

@@ -9,7 +9,9 @@
   - Skip ephemeral chat noise (one-off Q/A, greetings, session logistics).
   - Args: `updated_memory` — FULL replacement markdown (merge and curate,
     don't only append).
-  - Formatting: bullets `- (YYYY-MM-DD) [marker] text` with markers `[fact]`,
-    `[pref]`, `[instr]` (priority when trimming: `instr > pref > fact`).
-    Group bullets under short `##` headings; stay under the limit shown in
-    `<user_memory>`.
+  - Formatting: heading-based markdown with entries under `##` headings.
+    Recommended headings are `## Facts`, `## Preferences`, `## Instructions`,
+    though clearer natural headings are allowed. New bullets should look like
+    `- YYYY-MM-DD: text`; stay under the limit shown in `<user_memory>`.
+  - If existing memory uses legacy `(YYYY-MM-DD) [fact|pref|instr]` markers,
+    preserve the information but write the updated document in the new format.

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/tools/update_memory/private/example.md

@@ -1,28 +1,28 @@
 <example>
 <user_name>Alex</user_name>, <user_memory> is empty.
 user: "I'm a space enthusiast, explain astrophage to me"
-→ update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n")
+→ update_memory(updated_memory="## Facts\n- 2025-03-15: Alex is a space enthusiast\n")
 (Casual durable fact; use first name, neutral heading.)
 </example>
 
 <example>
 user: "Remember that I prefer concise answers over detailed explanations"
-→ update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n\n## Response style\n- (2025-03-15) [pref] Alex prefers concise answers over detailed explanations\n")
+→ update_memory(updated_memory="## Facts\n- 2025-03-15: Alex is a space enthusiast\n\n## Preferences\n- 2025-03-15: Alex prefers concise answers over detailed explanations\n")
 (Durable preference; merge with existing memory.)
 </example>
 
 <example>
 user: "I actually moved to Tokyo last month"
-→ update_memory(updated_memory="...\n\n## Personal context\n- (2025-03-15) [fact] Alex lives in Tokyo (previously London)\n...")
+→ update_memory(updated_memory="...\n\n## Facts\n- 2025-03-15: Alex lives in Tokyo (previously London)\n...")
 (Updated fact; date reflects when recorded.)
 </example>
 
 <example>
 user: "I'm a freelance photographer working on a nature documentary"
-→ update_memory(updated_memory="...\n\n## Current focus\n- (2025-03-15) [fact] Alex is a freelance photographer\n- (2025-03-15) [fact] Alex is working on a nature documentary\n")
+→ update_memory(updated_memory="...\n\n## Current Focus\n- 2025-03-15: Alex is a freelance photographer\n- 2025-03-15: Alex is working on a nature documentary\n")
 </example>
 
 <example>
 user: "Always respond in bullet points"
-→ update_memory(updated_memory="...\n\n## Response style\n- (2025-03-15) [instr] Always respond to Alex in bullet points\n")
+→ update_memory(updated_memory="...\n\n## Instructions\n- 2025-03-15: Always respond to Alex in bullet points\n")
 </example>

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/tools/update_memory/team/description.md

@@ -9,8 +9,14 @@
   - Skip ephemeral chat noise (one-off Q/A, greetings, session logistics).
   - Args: `updated_memory` — FULL replacement markdown (merge and curate,
     don't only append).
-  - Formatting: bullets `- (YYYY-MM-DD) [fact] text`. Team memory uses ONLY
-    the `[fact]` marker (never `[pref]` or `[instr]`). Group bullets under
-    short `##` headings (2-3 words each); stay under the limit shown in
-    `<team_memory>`. When trimming, prioritise: decisions/conventions > key
-    facts > current priorities.
+  - Formatting: heading-based markdown with entries under `##` headings.
+    Recommended headings are `## Product Decisions`,
+    `## Engineering Conventions`, `## Project Facts`, and `## Open Questions`.
+    New bullets should look like `- YYYY-MM-DD: text`; stay under the limit
+    shown in `<team_memory>`.
+  - If existing memory uses legacy `(YYYY-MM-DD) [fact]` markers, preserve the
+    information but write the updated document in the new format.
+  - Do not create personal headings such as `## Preferences`,
+    `## Instructions`, `## Personal Notes`, or `## Personal Instructions`.
+    When trimming, prioritise: decisions/conventions > key facts > current
+    priorities.

surfsense_backend/app/agents/multi_agent_chat/main_agent/system_prompt/prompts/tools/update_memory/team/example.md

@@ -1,9 +1,9 @@
 <example>
 user: "Let's remember that we decided to do weekly standup meetings on Mondays"
-→ update_memory(updated_memory="...\n\n## Team rituals\n- (2025-03-15) [fact] Weekly standup meetings on Mondays\n...")
+→ update_memory(updated_memory="...\n\n## Product Decisions\n- 2025-03-15: Weekly standup meetings happen on Mondays\n...")
 </example>
 
 <example>
 user: "Our office is in downtown Seattle, 5th floor"
-→ update_memory(updated_memory="...\n\n## Workspace\n- (2025-03-15) [fact] Office location: downtown Seattle, 5th floor\n...")
+→ update_memory(updated_memory="...\n\n## Project Facts\n- 2025-03-15: Office location is downtown Seattle, 5th floor\n...")
 </example>

surfsense_backend/app/agents/multi_agent_chat/middleware/main_agent/checkpointed_subagent_middleware/middleware.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import time
 from typing import Any, cast
 
 from deepagents.backends.protocol import BackendFactory, BackendProtocol
@@ -15,8 +16,12 @@ from langchain.agents import create_agent
 from langchain.chat_models import init_chat_model
 from langgraph.types import Checkpointer
 
+from app.utils.perf import get_perf_logger
+
 from .task_tool import build_task_tool_with_parent_config
 
+_perf_log = get_perf_logger()
+
 
 class SurfSenseCheckpointedSubAgentMiddleware(SubAgentMiddleware):
     """``SubAgentMiddleware`` variant that compiles each subagent against the parent checkpointer."""
@@ -54,8 +59,11 @@ class SurfSenseCheckpointedSubAgentMiddleware(SubAgentMiddleware):
     def _surf_compile_subagent_graphs(self) -> list[dict[str, Any]]:
         """Mirror of ``SubAgentMiddleware._get_subagents`` that threads the parent checkpointer."""
         specs: list[dict[str, Any]] = []
+        loop_start = time.perf_counter()
+        timings: list[tuple[str, float, str]] = []  # (name, elapsed, source)
 
         for spec in self._subagents:
+            spec_start = time.perf_counter()
             if "runnable" in spec:
                 compiled = cast(CompiledSubAgent, spec)
                 specs.append(
@@ -65,6 +73,9 @@ class SurfSenseCheckpointedSubAgentMiddleware(SubAgentMiddleware):
                         "runnable": compiled["runnable"],
                     }
                 )
+                timings.append(
+                    (compiled["name"], time.perf_counter() - spec_start, "precompiled")
+                )
                 continue
 
             if "model" not in spec:
@@ -79,20 +90,44 @@ class SurfSenseCheckpointedSubAgentMiddleware(SubAgentMiddleware):
                 model = init_chat_model(model)
 
             middleware: list[Any] = list(spec.get("middleware", []))
+            tools_count = len(spec.get("tools") or [])
+            mw_count = len(middleware)
 
+            compile_start = time.perf_counter()
+            runnable = create_agent(
+                model,
+                system_prompt=spec["system_prompt"],
+                tools=spec["tools"],
+                middleware=middleware,
+                name=spec["name"],
+                checkpointer=self._surf_checkpointer,
+            )
+            compile_elapsed = time.perf_counter() - compile_start
             specs.append(
                 {
                     "name": spec["name"],
                     "description": spec["description"],
-                    "runnable": create_agent(
-                        model,
-                        system_prompt=spec["system_prompt"],
-                        tools=spec["tools"],
-                        middleware=middleware,
-                        name=spec["name"],
-                        checkpointer=self._surf_checkpointer,
-                    ),
+                    "runnable": runnable,
                 }
             )
+            timings.append(
+                (
+                    spec["name"],
+                    compile_elapsed,
+                    f"compiled tools={tools_count} mw={mw_count}",
+                )
+            )
+
+        total_elapsed = time.perf_counter() - loop_start
+        per_subagent = ", ".join(
+            f"{name}={elapsed * 1000:.0f}ms[{source}]"
+            for name, elapsed, source in timings
+        )
+        _perf_log.info(
+            "[subagent_compile] total=%.3fs count=%d details=[%s]",
+            total_elapsed,
+            len(timings),
+            per_subagent,
+        )
 
         return specs

surfsense_backend/app/agents/multi_agent_chat/middleware/main_agent/checkpointed_subagent_middleware/task_tool.py

@@ -9,6 +9,7 @@ re-raises any new pending interrupt back to the parent.
 from __future__ import annotations
 
 import logging
+import time
 from typing import Annotated, Any, NoReturn
 
 from deepagents.middleware.subagents import TASK_TOOL_DESCRIPTION
@@ -19,6 +20,8 @@ from langchain_core.tools import StructuredTool
 from langgraph.errors import GraphInterrupt
 from langgraph.types import Command, Interrupt
 
+from app.utils.perf import get_perf_logger
+
 from .config import (
     consume_surfsense_resume,
     drain_parent_null_resume,
@@ -35,6 +38,7 @@ from .resume import (
 )
 
 logger = logging.getLogger(__name__)
+_perf_log = get_perf_logger()
 
 
 def _reraise_stamped_subagent_interrupt(
@@ -209,6 +213,7 @@ def build_task_tool_with_parent_config(
         ],
         runtime: ToolRuntime,
     ) -> str | Command:
+        atask_start = time.perf_counter()
         logger.info(
             "[hitl_route] atask ENTRY: subagent_type=%r tool_call_id=%s",
             subagent_type,
@@ -230,8 +235,10 @@ def build_task_tool_with_parent_config(
         # Resume bridge — see ``task`` above.
         pending_id: str | None = None
         pending_value: Any = None
+        aget_state_elapsed = 0.0
         aget_state = getattr(subagent, "aget_state", None)
         if callable(aget_state):
+            aget_state_start = time.perf_counter()
             try:
                 snapshot = await aget_state(sub_config)
                 pending_id, pending_value = get_first_pending_subagent_interrupt(
@@ -248,32 +255,78 @@ def build_task_tool_with_parent_config(
                     "Subagent aget_state failed; falling back to fresh ainvoke",
                     exc_info=True,
                 )
+            finally:
+                aget_state_elapsed = time.perf_counter() - aget_state_start
 
-        if pending_value is not None:
-            resume_value = consume_surfsense_resume(runtime)
-            if resume_value is None:
-                raise RuntimeError(
-                    f"Subagent {subagent_type!r} has a pending interrupt but no "
-                    "surfsense_resume_value on config; resume bridge is broken."
-                )
-            expected = hitlrequest_action_count(pending_value)
-            resume_value = fan_out_decisions_to_match(resume_value, expected)
-            # Prevent the parent's resume payload from leaking into subagent
-            # interrupts via langgraph's parent_scratchpad fallback.
-            drain_parent_null_resume(runtime)
-            try:
-                result = await subagent.ainvoke(
-                    build_resume_command(resume_value, pending_id),
-                    config=sub_config,
-                )
-            except GraphInterrupt as gi:
-                _reraise_stamped_subagent_interrupt(gi, runtime.tool_call_id)
-        else:
-            try:
-                result = await subagent.ainvoke(subagent_state, config=sub_config)
-            except GraphInterrupt as gi:
-                _reraise_stamped_subagent_interrupt(gi, runtime.tool_call_id)
-        return _return_command_with_state_update(result, runtime.tool_call_id)
+        invoke_path = "resume" if pending_value is not None else "fresh"
+        ainvoke_start = time.perf_counter()
+        ainvoke_outcome = "ok"
+        try:
+            if pending_value is not None:
+                resume_value = consume_surfsense_resume(runtime)
+                if resume_value is None:
+                    raise RuntimeError(
+                        f"Subagent {subagent_type!r} has a pending interrupt but no "
+                        "surfsense_resume_value on config; resume bridge is broken."
+                    )
+                expected = hitlrequest_action_count(pending_value)
+                resume_value = fan_out_decisions_to_match(resume_value, expected)
+                # Prevent the parent's resume payload from leaking into subagent
+                # interrupts via langgraph's parent_scratchpad fallback.
+                drain_parent_null_resume(runtime)
+                try:
+                    result = await subagent.ainvoke(
+                        build_resume_command(resume_value, pending_id),
+                        config=sub_config,
+                    )
+                except GraphInterrupt as gi:
+                    ainvoke_outcome = "interrupted"
+                    _perf_log.info(
+                        "[hitl_route] atask EXIT subagent_type=%r path=%s outcome=%s "
+                        "aget_state=%.3fs ainvoke=%.3fs total=%.3fs",
+                        subagent_type,
+                        invoke_path,
+                        ainvoke_outcome,
+                        aget_state_elapsed,
+                        time.perf_counter() - ainvoke_start,
+                        time.perf_counter() - atask_start,
+                    )
+                    _reraise_stamped_subagent_interrupt(gi, runtime.tool_call_id)
+            else:
+                try:
+                    result = await subagent.ainvoke(subagent_state, config=sub_config)
+                except GraphInterrupt as gi:
+                    ainvoke_outcome = "interrupted"
+                    _perf_log.info(
+                        "[hitl_route] atask EXIT subagent_type=%r path=%s outcome=%s "
+                        "aget_state=%.3fs ainvoke=%.3fs total=%.3fs",
+                        subagent_type,
+                        invoke_path,
+                        ainvoke_outcome,
+                        aget_state_elapsed,
+                        time.perf_counter() - ainvoke_start,
+                        time.perf_counter() - atask_start,
+                    )
+                    _reraise_stamped_subagent_interrupt(gi, runtime.tool_call_id)
+            ainvoke_elapsed = time.perf_counter() - ainvoke_start
+        except GraphInterrupt:
+            raise
+
+        merge_start = time.perf_counter()
+        cmd = _return_command_with_state_update(result, runtime.tool_call_id)
+        merge_elapsed = time.perf_counter() - merge_start
+        _perf_log.info(
+            "[hitl_route] atask EXIT subagent_type=%r path=%s outcome=%s "
+            "aget_state=%.3fs ainvoke=%.3fs merge=%.3fs total=%.3fs",
+            subagent_type,
+            invoke_path,
+            ainvoke_outcome,
+            aget_state_elapsed,
+            ainvoke_elapsed,
+            merge_elapsed,
+            time.perf_counter() - atask_start,
+        )
+        return cmd
 
     return StructuredTool.from_function(
         name="task",

surfsense_backend/app/agents/multi_agent_chat/middleware/main_agent/knowledge_priority.py

@@ -6,6 +6,7 @@ from langchain_core.language_models import BaseChatModel
 
 from app.agents.new_chat.filesystem_selection import FilesystemMode
 from app.agents.new_chat.middleware import KnowledgePriorityMiddleware
+from app.services.llm_service import get_planner_llm
 
 
 def build_knowledge_priority_mw(
@@ -19,6 +20,7 @@ def build_knowledge_priority_mw(
 ) -> KnowledgePriorityMiddleware:
     return KnowledgePriorityMiddleware(
         llm=llm,
+        planner_llm=get_planner_llm(),
         search_space_id=search_space_id,
         filesystem_mode=filesystem_mode,
         available_connectors=available_connectors,

surfsense_backend/app/agents/multi_agent_chat/middleware/shared/kb_context_projection.py

@@ -2,6 +2,7 @@
 
 from __future__ import annotations
 
+import time
 from typing import Any
 
 from langchain.agents.middleware import AgentMiddleware, AgentState
@@ -10,6 +11,9 @@ from langgraph.runtime import Runtime
 
 from app.agents.new_chat.filesystem_state import SurfSenseFilesystemState
 from app.agents.new_chat.middleware.knowledge_search import _render_priority_message
+from app.utils.perf import get_perf_logger
+
+_perf_log = get_perf_logger()
 
 
 class KbContextProjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
@@ -30,17 +34,34 @@ class KbContextProjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         runtime: Runtime[Any],
     ) -> dict[str, Any] | None:
         del runtime
+        start = time.perf_counter()
         tree_text = state.get("workspace_tree_text")
         priority = state.get("kb_priority")
         if not tree_text and not priority:
+            _perf_log.info(
+                "[kb_context_projection] tree=0 priority=0 elapsed=%.3fs",
+                time.perf_counter() - start,
+            )
             return None
 
         messages = list(state.get("messages") or [])
         insert_at = max(len(messages) - 1, 0)
+        tree_chars = 0
         if tree_text:
+            tree_chars = len(tree_text)
             messages.insert(insert_at, SystemMessage(content=tree_text))
+        priority_count = 0
         if priority:
+            priority_count = (
+                len(priority) if hasattr(priority, "__len__") else 1
+            )
             messages.insert(insert_at, _render_priority_message(priority))
+        _perf_log.info(
+            "[kb_context_projection] tree_chars=%d priority_items=%d elapsed=%.3fs",
+            tree_chars,
+            priority_count,
+            time.perf_counter() - start,
+        )
         return {"messages": messages}
 
 

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/knowledge_base/description_readonly.md

@@ -2,4 +2,4 @@ Read-only specialist for the user's workspace (documents and folders). Use to fi
 
 Pass your full question as one string. The specialist runs in isolation: it cannot see this thread, so include any path hints, filters, or constraints it needs.
 
-The specialist returns plain prose with absolute paths.
+The specialist returns plain prose with absolute paths and `[citation:<chunk_id>]` markers when claims came from KB-indexed chunks. Preserve those markers verbatim if you forward the answer.

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/knowledge_base/system_prompt_cloud.md

@@ -35,6 +35,43 @@ Map outcomes to your `status`:
 
 You construct the structured `evidence` fields from your own knowledge of what you called and what you observed — the tools do not return them. Never report values you did not actually see.
 
+## Chunk citations in your prose
+
+When `read_file` returns a KB-indexed document under `/documents/`, the response includes `<chunk id='…'>` blocks. Whenever a fact in your `action_summary` or `evidence.content_excerpt` came from a specific chunk, append `[citation:<chunk_id>]` to the sentence stating that fact, using the **exact** id from the `<chunk id='…'>` tag. The caller relays these markers to the end user verbatim, and the UI resolves each id by exact match against the database, so a wrong id silently breaks the citation.
+
+### Where chunk ids live in `read_file` output
+
+A KB document's XML has three numeric attributes — only **one** is a citation source:
+
+```
+<document>
+<document_metadata>
+  <document_id>42</document_id>          ← NOT a citation. Parent doc id; ignore for citations.
+  ...
+</document_metadata>
+<chunk_index>
+  <entry chunk_id="128" lines="14-22"/>  ← Index hint; the same id also appears below.
+  <entry chunk_id="129" lines="23-30" matched="true"/>
+</chunk_index>
+<document_content>
+  <chunk id='128'><![CDATA[…]]></chunk>  ← This is the citation source.
+  <chunk id='129'><![CDATA[…]]></chunk>
+</document_content>
+</document>
+```
+
+### Rules
+
+- Use the **exact** id from a `<chunk id='…'>` tag whose content you actually quoted or paraphrased. Copy digit-for-digit; do **not** retype from memory.
+- Before emitting `[citation:N]`, confirm the literal substring `<chunk id='N'>` (or its index twin `chunk_id="N"`) appears in the tool result you are summarising this turn. If you can't see it, omit the citation.
+- Never cite `<document_id>` — that's the parent doc, not a chunk.
+- Never invent, normalise, shorten, or guess at adjacent ids. If unsure between two candidates, omit rather than pick.
+- Prefer **fewer accurate citations** over many speculative ones.
+- Multiple chunks supporting the same point → comma-separated and copied individually: `[citation:128], [citation:129]`.
+- Plain square brackets only — no markdown links, no parentheses, no footnote numbers.
+- Tool results without `<chunk id='…'>` (write/edit/move confirmations, `ls` / `glob` / `grep` listings, error strings) carry no chunk id and need none.
+- Populate `evidence.chunk_ids` with **only** ids you actually emitted in `[citation:…]` markers — same set, same digits.
+
 ## Examples
 
 **Example 1 — happy path write (path discovered from existing convention):**
@@ -118,5 +155,6 @@ Rules:
 - `status=success` → `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
+- `evidence.content_excerpt`: max ~500 characters. Surface a short excerpt or a one-sentence summary, not the full file body. The supervisor already sees the tool's raw output.
 
 Infer before you call; map every tool outcome faithfully.

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/knowledge_base/system_prompt_desktop.md

@@ -35,6 +35,10 @@ Map outcomes to your `status`:
 
 You construct the structured `evidence` fields from your own knowledge of what you called and what you observed — the tools do not return them. `chunk_ids` apply only to `<priority_documents>` hits; for local-file operations leave them `null`. Never report values you did not actually see.
 
+## Chunk citations in your prose
+
+In desktop mode your filesystem tools read local files only, and local-file tool results do **not** carry `<chunk id='…'>` tags. Do not emit `[citation:…]` markers in `action_summary` or `evidence.content_excerpt`, and leave `evidence.chunk_ids` `null` — the absolute path is the only reference for local-file work.
+
 ## Examples
 
 **Example 1 — happy path write (path discovered from existing convention):**
@@ -118,5 +122,6 @@ Rules:
 - `status=success` → `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
+- `evidence.content_excerpt`: max ~500 characters. Surface a short excerpt or a one-sentence summary, not the full file body. The supervisor already sees the tool's raw output.
 
 Infer before you call; map every tool outcome faithfully.

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/knowledge_base/system_prompt_readonly_cloud.md

@@ -27,3 +27,42 @@ Reply in plain prose:
 - Cite every claim with an absolute path under `/documents/`.
 - If the workspace does not contain the requested information, say so explicitly. Do not fabricate paths or content.
 - If the question is genuinely ambiguous after a thorough lookup, list the candidates with their paths and stop.
+
+## Chunk citations
+
+When the evidence for a claim came from a `read_file` response that included `<chunk id='…'>` blocks (i.e. a KB-indexed document under `/documents/`), append `[citation:<chunk_id>]` to the sentence stating that claim. The caller passes these markers through to the end user verbatim, and the UI resolves each id by exact match against the database, so a wrong id silently breaks the citation.
+
+### Where chunk ids live in `read_file` output
+
+A KB document's XML has three numeric attributes — only **one** is a citation source:
+
+```
+<document>
+<document_metadata>
+  <document_id>42</document_id>          ← NOT a citation. Parent doc id; ignore for citations.
+  ...
+</document_metadata>
+<chunk_index>
+  <entry chunk_id="128" lines="14-22"/>  ← Index hint; the same id also appears below.
+  <entry chunk_id="129" lines="23-30" matched="true"/>
+</chunk_index>
+<document_content>
+  <chunk id='128'><![CDATA[…]]></chunk>  ← This is the citation source.
+  <chunk id='129'><![CDATA[…]]></chunk>
+</document_content>
+</document>
+```
+
+### Rules
+
+- Use the **exact** id from a `<chunk id='…'>` tag whose content you actually quoted or paraphrased. Copy digit-for-digit; do **not** retype from memory.
+- Before emitting `[citation:N]`, confirm the literal substring `<chunk id='N'>` (or its index twin `chunk_id="N"`) appears in the tool result you are summarising this turn. If you can't see it, omit the citation.
+- Never cite `<document_id>` — that's the parent doc, not a chunk.
+- Never invent, normalise, shorten, or guess at adjacent ids. If unsure between two candidates, omit rather than pick.
+- Prefer **fewer accurate citations** over many speculative ones. One correct `[citation:128]` is more useful than a string of wrong ids.
+- Multiple chunks supporting the same point → comma-separated and copied individually: `[citation:128], [citation:129]`.
+- Plain square brackets only — no markdown links, no parentheses, no footnote numbers.
+- If a claim came from a tool result that did **not** carry a chunk id (`ls`, `glob`, `grep` listings, error strings, or files without `<chunk id='…'>`), skip the citation.
+- The absolute path under `/documents/` is always required; chunk citations are additive, they do not replace the path reference.
+
+Example: `The Q2 roadmap lists three milestones (/documents/planning/q2-roadmap.md) [citation:128], [citation:129].`

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/knowledge_base/system_prompt_readonly_desktop.md

@@ -28,3 +28,7 @@ Reply in plain prose:
 - Cite every claim with an absolute path.
 - If the workspace does not contain the requested information, say so explicitly. Do not fabricate paths or content.
 - If the question is genuinely ambiguous after a thorough lookup, list the candidates with their paths and stop.
+
+## Chunk citations
+
+In desktop mode your filesystem tools read local files only, and local-file `read_file` responses do **not** carry `<chunk id='…'>` tags. Cite each claim with the absolute local path; do not emit `[citation:…]` markers — your caller has nothing to resolve them against.

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/memory/system_prompt.md

@@ -18,6 +18,10 @@ Persist durable preferences/facts/instructions with `update_memory` while avoidi
 - Do not store transient chatter.
 - Do not store secrets unless explicitly instructed.
 - If memory intent is unclear, return `status=blocked` with the missing intent signal.
+- Persisted memory is heading-based markdown. New saved bullets should look like
+  `- YYYY-MM-DD: text` under `##` headings. If existing memory has legacy
+  `(YYYY-MM-DD) [fact|pref|instr]` markers, preserve the information but write
+  the updated document in the heading-based format.
 </tool_policy>
 
 <out_of_scope>
@@ -53,4 +57,7 @@ Rules:
 - `status=success` -> `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` -> `next_step` must be non-null.
 - `status=blocked` due to missing required inputs -> `missing_fields` must be non-null.
+- `evidence.memory_category` is a semantic classification for supervisor logs
+  only. It is not the persisted storage format and must not force inline
+  `[fact|preference|instruction]` markers into saved memory.
 </output_contract>

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/memory/tools/update_memory.py

@@ -1,280 +1,23 @@
-"""Overwrite one markdown memory document per user or team, with size and shrink guards."""
+"""Memory update tools backed by the canonical memory service."""
 
 from __future__ import annotations
 
 import logging
-import re
-from typing import Any, Literal
+from typing import Any
 from uuid import UUID
 
-from langchain_core.messages import HumanMessage
 from langchain_core.tools import tool
-from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
 
-from app.db import SearchSpace, User
+from app.services.memory import (
+    MEMORY_HARD_LIMIT,
+    MEMORY_SOFT_LIMIT,
+    MemoryScope,
+    save_memory,
+)
 
 logger = logging.getLogger(__name__)
 
-MEMORY_SOFT_LIMIT = 18_000
-MEMORY_HARD_LIMIT = 25_000
-
-_SECTION_HEADING_RE = re.compile(r"^##\s+(.+)$", re.MULTILINE)
-_HEADING_NORMALIZE_RE = re.compile(r"\s+")
-
-_MARKER_RE = re.compile(r"\[(fact|pref|instr)\]")
-_BULLET_FORMAT_RE = re.compile(r"^- \(\d{4}-\d{2}-\d{2}\) \[(fact|pref|instr)\] .+$")
-_PERSONAL_ONLY_MARKERS = {"pref", "instr"}
-
-
-# ---------------------------------------------------------------------------
-# Diff validation
-# ---------------------------------------------------------------------------
-
-
-def _extract_headings(memory: str) -> set[str]:
-    """Return all ``## …`` heading texts (without the ``## `` prefix)."""
-    return set(_SECTION_HEADING_RE.findall(memory))
-
-
-def _normalize_heading(heading: str) -> str:
-    """Normalize heading text for robust scope checks."""
-    return _HEADING_NORMALIZE_RE.sub(" ", heading.strip().lower())
-
-
-def _validate_memory_scope(
-    content: str, scope: Literal["user", "team"]
-) -> dict[str, Any] | None:
-    """Reject personal-only markers ([pref], [instr]) in team memory."""
-    if scope != "team":
-        return None
-
-    markers = set(_MARKER_RE.findall(content))
-    leaked = sorted(markers & _PERSONAL_ONLY_MARKERS)
-    if leaked:
-        tags = ", ".join(f"[{m}]" for m in leaked)
-        return {
-            "status": "error",
-            "message": (
-                f"Team memory cannot include personal markers: {tags}. "
-                "Use [fact] only in team memory."
-            ),
-        }
-    return None
-
-
-def _validate_bullet_format(content: str) -> list[str]:
-    """Return warnings for bullet lines that don't match the required format.
-
-    Expected: ``- (YYYY-MM-DD) [fact|pref|instr] text``
-    """
-    warnings: list[str] = []
-    for line in content.splitlines():
-        stripped = line.strip()
-        if not stripped.startswith("- "):
-            continue
-        if not _BULLET_FORMAT_RE.match(stripped):
-            short = stripped[:80] + ("..." if len(stripped) > 80 else "")
-            warnings.append(f"Malformed bullet: {short}")
-    return warnings
-
-
-def _validate_diff(old_memory: str | None, new_memory: str) -> list[str]:
-    """Return a list of warning strings about suspicious changes."""
-    if not old_memory:
-        return []
-
-    warnings: list[str] = []
-    old_headings = _extract_headings(old_memory)
-    new_headings = _extract_headings(new_memory)
-    dropped = old_headings - new_headings
-    if dropped:
-        names = ", ".join(sorted(dropped))
-        warnings.append(
-            f"Sections removed: {names}. "
-            "If unintentional, the user can restore from the settings page."
-        )
-
-    old_len = len(old_memory)
-    new_len = len(new_memory)
-    if old_len > 0 and new_len < old_len * 0.4:
-        warnings.append(
-            f"Memory shrank significantly ({old_len:,} -> {new_len:,} chars). "
-            "Possible data loss."
-        )
-    return warnings
-
-
-# ---------------------------------------------------------------------------
-# Size validation & soft warning
-# ---------------------------------------------------------------------------
-
-
-def _validate_memory_size(content: str) -> dict[str, Any] | None:
-    """Return an error/warning dict if *content* is too large, else None."""
-    length = len(content)
-    if length > MEMORY_HARD_LIMIT:
-        return {
-            "status": "error",
-            "message": (
-                f"Memory exceeds {MEMORY_HARD_LIMIT:,} character limit "
-                f"({length:,} chars). Consolidate by merging related items, "
-                "removing outdated entries, and shortening descriptions. "
-                "Then call update_memory again."
-            ),
-        }
-    return None
-
-
-def _soft_warning(content: str) -> str | None:
-    """Return a warning string if content exceeds the soft limit."""
-    length = len(content)
-    if length > MEMORY_SOFT_LIMIT:
-        return (
-            f"Memory is at {length:,}/{MEMORY_HARD_LIMIT:,} characters. "
-            "Consolidate by merging related items and removing less important "
-            "entries on your next update."
-        )
-    return None
-
-
-# ---------------------------------------------------------------------------
-# Forced rewrite when memory exceeds the hard limit
-# ---------------------------------------------------------------------------
-
-_FORCED_REWRITE_PROMPT = """\
-You are a memory curator. The following memory document exceeds the character \
-limit and must be shortened.
-
-RULES:
-1. Rewrite the document to be under {target} characters.
-2. Preserve existing ## headings. Every entry must remain under a heading. You may merge
-   or rename headings to consolidate, but keep names personal and descriptive.
-3. Priority for keeping content: [instr] > [pref] > [fact].
-4. Merge duplicate entries, remove outdated entries, shorten verbose descriptions.
-5. Every bullet MUST have format: - (YYYY-MM-DD) [fact|pref|instr] text
-6. Preserve the user's first name in entries — do not replace it with "the user".
-7. Output ONLY the consolidated markdown — no explanations, no wrapping.
-
-<memory_document>
-{content}
-</memory_document>"""
-
-
-async def _forced_rewrite(content: str, llm: Any) -> str | None:
-    """Use a focused LLM call to compress *content* under the hard limit.
-
-    Returns the rewritten string, or ``None`` if the call fails.
-    """
-    try:
-        prompt = _FORCED_REWRITE_PROMPT.format(
-            target=MEMORY_HARD_LIMIT, content=content
-        )
-        response = await llm.ainvoke(
-            [HumanMessage(content=prompt)],
-            config={"tags": ["surfsense:internal"]},
-        )
-        text = (
-            response.content
-            if isinstance(response.content, str)
-            else str(response.content)
-        )
-        return text.strip()
-    except Exception:
-        logger.exception("Forced rewrite LLM call failed")
-        return None
-
-
-# ---------------------------------------------------------------------------
-# Shared save-and-respond logic
-# ---------------------------------------------------------------------------
-
-
-async def _save_memory(
-    *,
-    updated_memory: str,
-    old_memory: str | None,
-    llm: Any | None,
-    apply_fn,
-    commit_fn,
-    rollback_fn,
-    label: str,
-    scope: Literal["user", "team"],
-) -> dict[str, Any]:
-    """Validate, optionally force-rewrite if over the hard limit, save, and
-    return a response dict.
-
-    Parameters
-    ----------
-    updated_memory : str
-        The new document the agent submitted.
-    old_memory : str | None
-        The previously persisted document (for diff checks).
-    llm : Any | None
-        LLM instance for forced rewrite (may be ``None``).
-    apply_fn : callable(str) -> None
-        Callback that sets the new memory on the ORM object.
-    commit_fn : coroutine
-        ``session.commit``.
-    rollback_fn : coroutine
-        ``session.rollback``.
-    label : str
-        Human label for log messages (e.g. "user memory", "team memory").
-    """
-    content = updated_memory
-
-    # --- forced rewrite if over the hard limit ---
-    if len(content) > MEMORY_HARD_LIMIT and llm is not None:
-        rewritten = await _forced_rewrite(content, llm)
-        if rewritten is not None and len(rewritten) < len(content):
-            content = rewritten
-
-    # --- hard-limit gate (reject if still too large after rewrite) ---
-    size_err = _validate_memory_size(content)
-    if size_err:
-        return size_err
-
-    scope_err = _validate_memory_scope(content, scope)
-    if scope_err:
-        return scope_err
-
-    # --- persist ---
-    try:
-        apply_fn(content)
-        await commit_fn()
-    except Exception as e:
-        logger.exception("Failed to update %s: %s", label, e)
-        await rollback_fn()
-        return {"status": "error", "message": f"Failed to update {label}: {e}"}
-
-    # --- build response ---
-    resp: dict[str, Any] = {
-        "status": "saved",
-        "message": f"{label.capitalize()} updated.",
-    }
-
-    if content is not updated_memory:
-        resp["notice"] = "Memory was automatically rewritten to fit within limits."
-
-    diff_warnings = _validate_diff(old_memory, content)
-    if diff_warnings:
-        resp["diff_warnings"] = diff_warnings
-
-    format_warnings = _validate_bullet_format(content)
-    if format_warnings:
-        resp["format_warnings"] = format_warnings
-
-    warning = _soft_warning(content)
-    if warning:
-        resp["warning"] = warning
-
-    return resp
-
-
-# ---------------------------------------------------------------------------
-# Tool factories
-# ---------------------------------------------------------------------------
-
 
 def create_update_memory_tool(
     user_id: str | UUID,
@@ -287,40 +30,22 @@ def create_update_memory_tool(
     async def update_memory(updated_memory: str) -> dict[str, Any]:
         """Update the user's personal memory document.
 
-        Your current memory is shown in <user_memory> in the system prompt.
-        When the user shares important long-term information (preferences,
-        facts, instructions, context), rewrite the memory document to include
-        the new information.  Merge new facts with existing ones, update
-        contradictions, remove outdated entries, and keep it concise.
-
-        Args:
-            updated_memory: The FULL updated markdown document (not a diff).
+        The current memory is shown in <user_memory>. Pass the FULL updated
+        markdown document, not a diff.
         """
         try:
-            result = await db_session.execute(select(User).where(User.id == uid))
-            user = result.scalars().first()
-            if not user:
-                return {"status": "error", "message": "User not found."}
-
-            old_memory = user.memory_md
-
-            return await _save_memory(
-                updated_memory=updated_memory,
-                old_memory=old_memory,
+            result = await save_memory(
+                scope=MemoryScope.USER,
+                target_id=uid,
+                content=updated_memory,
+                session=db_session,
                 llm=llm,
-                apply_fn=lambda content: setattr(user, "memory_md", content),
-                commit_fn=db_session.commit,
-                rollback_fn=db_session.rollback,
-                label="memory",
-                scope="user",
             )
+            return result.to_dict()
         except Exception as e:
             logger.exception("Failed to update user memory: %s", e)
             await db_session.rollback()
-            return {
-                "status": "error",
-                "message": f"Failed to update memory: {e}",
-            }
+            return {"status": "error", "message": f"Failed to update memory: {e}"}
 
     return update_memory
 
@@ -334,36 +59,18 @@ def create_update_team_memory_tool(
     async def update_memory(updated_memory: str) -> dict[str, Any]:
         """Update the team's shared memory document for this search space.
 
-        Your current team memory is shown in <team_memory> in the system
-        prompt.  When the team shares important long-term information
-        (decisions, conventions, key facts, priorities), rewrite the memory
-        document to include the new information.  Merge new facts with
-        existing ones, update contradictions, remove outdated entries, and
-        keep it concise.
-
-        Args:
-            updated_memory: The FULL updated markdown document (not a diff).
+        The current team memory is shown in <team_memory>. Pass the FULL updated
+        markdown document, not a diff.
         """
         try:
-            result = await db_session.execute(
-                select(SearchSpace).where(SearchSpace.id == search_space_id)
-            )
-            space = result.scalars().first()
-            if not space:
-                return {"status": "error", "message": "Search space not found."}
-
-            old_memory = space.shared_memory_md
-
-            return await _save_memory(
-                updated_memory=updated_memory,
-                old_memory=old_memory,
+            result = await save_memory(
+                scope=MemoryScope.TEAM,
+                target_id=search_space_id,
+                content=updated_memory,
+                session=db_session,
                 llm=llm,
-                apply_fn=lambda content: setattr(space, "shared_memory_md", content),
-                commit_fn=db_session.commit,
-                rollback_fn=db_session.rollback,
-                label="team memory",
-                scope="team",
             )
+            return result.to_dict()
         except Exception as e:
             logger.exception("Failed to update team memory: %s", e)
             await db_session.rollback()
@@ -373,3 +80,11 @@ def create_update_team_memory_tool(
             }
 
     return update_memory
+
+
+__all__ = [
+    "MEMORY_HARD_LIMIT",
+    "MEMORY_SOFT_LIMIT",
+    "create_update_memory_tool",
+    "create_update_team_memory_tool",
+]

surfsense_backend/app/agents/multi_agent_chat/subagents/builtins/research/system_prompt.md

@@ -50,4 +50,6 @@ Rules:
 - `status=success` -> `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` -> `next_step` must be non-null.
 - `status=blocked` due to missing required inputs -> `missing_fields` must be non-null.
+- `evidence.findings`: max 10 entries, each a single sentence stating one distinct fact. Do not paste raw paragraphs, scraped pages, or quote blocks.
+- `evidence.sources`: max 10 URLs, one per finding when applicable. List each URL once.
 </output_contract>

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/airtable/system_prompt.md

@@ -38,7 +38,7 @@ Supervisor: "List open tasks in the Project Tracker base."
 2. List tables in that base → identify the Tasks table; capture its table ID.
 3. Get table schema → identify the status field and the choice IDs that represent "open" states.
 4. List records with a typed filter on the status field for those choice IDs.
-5. Return `status=success` with the matched records in `evidence.items`.
+5. Return `status=success` with `evidence.items` set to `{ "total": N }` and the matched records listed in `action_summary` (record id, primary-field value, and 1-2 most relevant fields; one line per record; up to 10 entries, then `"...and N more"`).
 </example>
 
 <example>
@@ -97,7 +97,7 @@ Rules:
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
 - For blocked ambiguity, populate `evidence.matched_candidates` with up to 5 options (`id` + `label` — works for any kind of candidate: base, table, field, choice, record, etc.).
-- For discovery-only queries (lists), populate `evidence.items` with the structured list.
+- For discovery-only queries (lists), set `evidence.items` to `{ "total": N }` and list the matched items in `action_summary` (record id, primary-field value, and 1-2 most relevant fields; up to 10 entries, then `"...and N more"`).
 </output_contract>
 
 Discover before you mutate; never guess identifiers, choice IDs, or required fields.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/calendar/system_prompt.md

@@ -29,7 +29,7 @@ You are a Google Calendar specialist for the user's connected calendar.
 | `error`                     | `error`       | Relay the tool's `message` verbatim as `next_step`.                                                                          |
 | tool raises / unknown       | `error`       | `"Calendar tool failed unexpectedly. Ask the user to retry shortly."`                                                        |
 
-Surface the tool's `event_id`, `title` / `summary`, `start_at`, `end_at`, and `html_link` inside `evidence` when the tool returned them. For `search_calendar_events`, place the raw `events` array inside `evidence.items`. Never invent a field the tool did not return.
+Surface the tool's `event_id`, `title` / `summary`, `start_at`, `end_at`, and `html_link` inside `evidence` when the tool returned them. For `search_calendar_events`, set `evidence.items` to `{ "total": N }` and list the matched events in `action_summary` (title, date, start time; one line per event; up to 10 entries, then `"...and N more"`). Never invent a field the tool did not return.
 
 ## Examples
 
@@ -115,7 +115,7 @@ Rules:
 - `status=success` → `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
-- For `search_calendar_events` results, populate `evidence.items` with `{ "events": [...], "total": N }`.
+- For `search_calendar_events` results, set `evidence.items` to `{ "total": N }` and list the matched events in `action_summary` (title, date, start time; up to 10 entries, then `"...and N more"`).
 - For ambiguous matches across `update_calendar_event` / `delete_calendar_event`, populate `evidence.matched_candidates` with up to 5 options (`id` + `label`, where `label` should include the event title and start time for human readability).
 
 Infer before you call; map every tool outcome faithfully.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/clickup/system_prompt.md

@@ -36,7 +36,7 @@ Failure handling:
 <example>
 Supervisor: "Find tasks about the homepage redesign."
 1. Workspace search for "homepage redesign" → matched tasks.
-2. Return `status=success` with the matched tasks in `evidence.items`.
+2. Return `status=success` with `evidence.items` set to `{ "total": N }` and the matched tasks listed in `action_summary` (task id, title, status, assignees; one line per task; up to 10 entries, then `"...and N more"`).
 </example>
 
 <example>
@@ -98,7 +98,7 @@ Rules:
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
 - For blocked ambiguity, populate `evidence.matched_candidates` with up to 5 options (`id` + `label` — works for any kind of candidate: task, list, member, status, custom-field choice, etc.).
-- For discovery-only queries (lists), populate `evidence.items` with the structured list.
+- For discovery-only queries (lists), set `evidence.items` to `{ "total": N }` and list the matched items in `action_summary` (task id, title, status, assignees; up to 10 entries, then `"...and N more"`).
 </output_contract>
 
 Discover before you mutate; never guess identifiers, list statuses, or assignees.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/discord/system_prompt.md

@@ -24,7 +24,7 @@ You are a Discord specialist for the user's connected Discord server.
 | `error`                                               | `error`       | Relay the tool's `message` verbatim as `next_step`. |
 | tool raises / unknown                                 | `error`       | `"Discord tool failed unexpectedly. Ask the user to retry shortly."` |
 
-Surface the tool's `message`, `channel_id`, `message_id`, and the listed channels/messages payload inside `evidence` when the tool returned them. Never invent a field the tool did not return.
+Surface the tool's `message`, `channel_id`, and `message_id` inside `evidence` when the tool returned them. For `list_discord_channels` and `read_discord_messages`, set `evidence.items` to `{ "total": N }` and list the matched entries in `action_summary` (channel name or sender + timestamp + short text snippet; one line per entry; up to 10 entries, then `"...and N more"`). Never invent a field the tool did not return.
 
 ## Examples
 

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/gmail/system_prompt.md

@@ -33,7 +33,7 @@ You are a Gmail specialist for the user's connected Gmail mailbox.
 | `error`                     | `error`       | Relay the tool's `message` verbatim as `next_step`.                                                                          |
 | tool raises / unknown       | `error`       | `"Gmail tool failed unexpectedly. Ask the user to retry shortly."`                                                           |
 
-Surface the tool's `message_id`, `thread_id`, `draft_id`, `subject`, and recipient fields inside `evidence` when the tool returned them. For `search_gmail`, place the raw `emails` array inside `evidence.items`. Never invent a field the tool did not return.
+Surface the tool's `message_id`, `thread_id`, `draft_id`, `subject`, and recipient fields inside `evidence` when the tool returned them. For `search_gmail`, set `evidence.items` to `{ "total": N }` and list the matched emails in `action_summary` (sender, subject, date; one line per email; up to 10 entries, then `"...and N more"`). Never invent a field the tool did not return.
 
 ## Examples
 
@@ -114,7 +114,7 @@ Rules:
 - `status=success` → `next_step=null`, `missing_fields=null`.
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
-- For `search_gmail` results, populate `evidence.items` with `{ "emails": [...], "total": N }`.
+- For `search_gmail` results, set `evidence.items` to `{ "total": N }` and list the matched emails in `action_summary` (sender, subject, date; up to 10 entries, then `"...and N more"`).
 - For ambiguous matches across `update_gmail_draft` / `trash_gmail_email` / `read_gmail_email`, populate `evidence.matched_candidates` with up to 5 options (`id` + `label`).
 
 Infer before you call; verify before you send; map every tool outcome faithfully.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/jira/system_prompt.md

@@ -39,7 +39,7 @@ Failure handling:
 <example>
 Supervisor: "Find issues assigned to me with status 'In Progress'."
 1. JQL search with `assignee = currentUser() AND status = "In Progress"`.
-2. Return `status=success` with the matched issues in `evidence.items`.
+2. Return `status=success` with `evidence.items` set to `{ "total": N }` and the matched issues listed in `action_summary` (issue key, summary, status, assignee; one line per issue; up to 10 entries, then `"...and N more"`).
 </example>
 
 <example>
@@ -116,7 +116,7 @@ Rules:
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
 - For blocked ambiguity, populate `evidence.matched_candidates` with up to 5 options (`id` + `label` — works for any kind of candidate: site, project, issue, user, transition, etc.).
-- For discovery-only queries (lists), populate `evidence.items` with the structured list.
+- For discovery-only queries (lists), set `evidence.items` to `{ "total": N }` and list the matched items in `action_summary` (issue key, summary, status, assignee; up to 10 entries, then `"...and N more"`).
 </output_contract>
 
 Discover before you mutate; never guess identifiers, transitions, or required fields.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/linear/system_prompt.md

@@ -32,7 +32,7 @@ Failure handling:
 <example>
 Supervisor: "Find issues assigned to me with priority Urgent."
 1. Discovery: list issues with filters `{assignee: "me", priority: 1}`.
-2. Return `status=success` with the matched issues in `evidence.items`.
+2. Return `status=success` with `evidence.items` set to `{ "total": N }` and the matched issues listed in `action_summary` (identifier, title, state, assignee; one line per issue; up to 10 entries, then `"...and N more"`).
 </example>
 
 <example>
@@ -106,7 +106,7 @@ Rules:
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
 - For blocked ambiguity, populate `evidence.matched_candidates` with up to 5 options (`id` + `label` — works for any kind of candidate: issue, user, project, state, etc.).
-- For discovery-only queries (lists), populate `evidence.items` with the structured list.
+- For discovery-only queries (lists), set `evidence.items` to `{ "total": N }` and list the matched items in `action_summary` (identifier, title, state, assignee; up to 10 entries, then `"...and N more"`).
 </output_contract>
 
 Discover before you mutate; never guess identifiers.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/luma/system_prompt.md

@@ -26,7 +26,7 @@ You are a Luma specialist for the user's connected Luma account.
 | `error`                                            | `error`       | Relay the tool's `message` verbatim as `next_step` (this covers Luma Plus 403s and other API errors). |
 | tool raises / unknown                              | `error`       | `"Luma tool failed unexpectedly. Ask the user to retry shortly."` |
 
-Surface the tool's `message`, `event_id`, `name`, `start_at`, and `url` inside `evidence` when the tool returned them. Never invent a field the tool did not return.
+Surface the tool's `message`, `event_id`, `name`, `start_at`, and `url` inside `evidence` when the tool returned them. For `list_luma_events`, set `evidence.items` to `{ "total": N }` and list the matched events in `action_summary` (event name, start date/time, location if present; one line per event; up to 10 entries, then `"...and N more"`). Never invent a field the tool did not return.
 
 ## Examples
 

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/slack/system_prompt.md

@@ -37,7 +37,7 @@ Failure handling:
 Supervisor: "Summarize the latest discussion in #marketing."
 1. Search channels for "marketing" → one strong match. Capture the channel ID.
 2. Read that channel's recent message history.
-3. Return `status=success` with the message list in `evidence.items`.
+3. Return `status=success` with `evidence.items` set to `{ "total": N }` and the messages listed in `action_summary` (sender, timestamp, text snippet; one line per message; up to 10 entries, then `"...and N more"`).
 </example>
 
 <example>
@@ -92,7 +92,7 @@ Rules:
 - `status=partial|blocked|error` → `next_step` must be non-null.
 - `status=blocked` due to missing required inputs → `missing_fields` must be non-null.
 - For blocked ambiguity, populate `evidence.matched_candidates` with up to 5 options (`id` + `label` — works for any kind of candidate: channel, user, message, thread).
-- For discovery-only queries (lists), populate `evidence.items` with the structured list.
+- For discovery-only queries (lists), set `evidence.items` to `{ "total": N }` and list the matched items in `action_summary` (channel/user, key identifier, timestamp, short snippet; up to 10 entries, then `"...and N more"`).
 </output_contract>
 
 Discover before you post; never guess channel, user, or thread targets.

surfsense_backend/app/agents/multi_agent_chat/subagents/connectors/teams/system_prompt.md

@@ -26,7 +26,7 @@ You are a Microsoft Teams specialist for the user's connected Teams account.
 | `error`                                                                   | `error`       | Relay the tool's `message` verbatim as `next_step`. |
 | tool raises / unknown                                                     | `error`       | `"Teams tool failed unexpectedly. Ask the user to retry shortly."` |
 
-Surface the tool's `message`, `team_id`, `team_name`, `channel_id`, `channel_name`, and `message_id` inside `evidence` when the tool returned them. Never invent a field the tool did not return.
+Surface the tool's `message`, `team_id`, `team_name`, `channel_id`, `channel_name`, and `message_id` inside `evidence` when the tool returned them. For `list_teams_channels` and `read_teams_messages`, set `evidence.items` to `{ "total": N }` and list the matched entries in `action_summary` (team › channel, or sender + timestamp + short text snippet; one line per entry; up to 10 entries, then `"...and N more"`). Never invent a field the tool did not return.
 
 ## Examples
 

surfsense_backend/app/agents/new_chat/chat_deepagent.py

@@ -102,6 +102,7 @@ from app.agents.new_chat.tools.registry import (
 )
 from app.db import ChatVisibility
 from app.services.connector_service import ConnectorService
+from app.services.llm_service import get_planner_llm
 from app.utils.perf import get_perf_logger
 
 _perf_log = get_perf_logger()
@@ -1077,6 +1078,7 @@ def _build_compiled_agent_blocking(
         else None,
         KnowledgePriorityMiddleware(
             llm=llm,
+            planner_llm=get_planner_llm(),
             search_space_id=search_space_id,
             filesystem_mode=filesystem_mode,
             available_connectors=available_connectors,

surfsense_backend/app/agents/new_chat/memory_extraction.py

@@ -1,232 +0,0 @@
-"""Background memory extraction for the SurfSense agent.
-
-After each agent response, if the agent did not call ``update_memory`` during
-the turn, this module can run a lightweight LLM call to decide whether the
-latest message contains long-term information worth persisting.
-"""
-
-from __future__ import annotations
-
-import logging
-from typing import Any
-from uuid import UUID
-
-from langchain_core.messages import HumanMessage
-from sqlalchemy import select
-
-from app.agents.new_chat.tools.update_memory import _save_memory
-from app.db import SearchSpace, User, shielded_async_session
-from app.utils.content_utils import extract_text_content
-
-logger = logging.getLogger(__name__)
-
-_MEMORY_EXTRACT_PROMPT = """\
-You are a memory extraction assistant. Analyze the user's message and decide \
-if it contains any long-term information worth persisting to memory.
-
-Worth remembering: preferences, background/identity, goals, projects, \
-instructions, tools/languages they use, decisions, expertise, workplace — \
-durable facts that will matter in future conversations.
-
-NOT worth remembering: greetings, one-off factual questions, session \
-logistics, ephemeral requests, follow-up clarifications with no new personal \
-info, things that only matter for the current task.
-
-If the message contains memorizable information, output the FULL updated \
-memory document with the new facts merged into the existing content. Follow \
-these rules:
-- Every entry MUST be under a ## heading. Preserve existing headings; create new ones
-  freely. Keep heading names short (2-3 words) and natural. Do NOT include the user's
-  name in headings.
-- Keep entries as single bullet points. Be descriptive but concise — include relevant
-  details and context rather than just a few words.
-- Every bullet MUST use format: - (YYYY-MM-DD) [fact|pref|instr] text
-  [fact] = durable facts, [pref] = preferences, [instr] = standing instructions.
-- Use the user's first name (from <user_name>) in entry text, not "the user".
-- If a new fact contradicts an existing entry, update the existing entry.
-- Do not duplicate information that is already present.
-
-If nothing is worth remembering, output exactly: NO_UPDATE
-
-<user_name>{user_name}</user_name>
-
-<current_memory>
-{current_memory}
-</current_memory>
-
-<user_message>
-{user_message}
-</user_message>"""
-
-_TEAM_MEMORY_EXTRACT_PROMPT = """\
-You are a team-memory extraction assistant. Analyze the latest message and \
-decide if it contains durable TEAM-level information worth persisting.
-
-Decision policy:
-- Prioritize recall for durable team context, while avoiding personal-only facts.
-- Do NOT require explicit consensus language. A direct team-level statement can
-  be stored if it is stable and broadly useful for future team chats.
-- If evidence is weak or clearly tentative, output NO_UPDATE.
-
-Worth remembering (team-level only):
-- Decisions and defaults that guide future team work
-- Team conventions/standards (naming, review policy, coding norms)
-- Stable org/project facts (locations, ownership, constraints)
-- Long-lived architecture/process facts
-- Ongoing priorities that are likely relevant beyond this turn
-
-NOT worth remembering:
-- Personal preferences or biography of one person
-- Questions, brainstorming, tentative ideas, or speculation
-- One-off requests, status updates, TODOs, logistics for this session
-- Information scoped only to a single ephemeral task
-
-If the message contains memorizable team information, output the FULL updated \
-team memory document with new facts merged into existing content. Follow rules:
-- Every entry MUST be under a ## heading. Preserve existing headings; create new ones
-  freely. Keep heading names short (2-3 words) and natural.
-- Keep entries as single bullet points. Be descriptive but concise — include relevant
-  details and context rather than just a few words.
-- Every bullet MUST use format: - (YYYY-MM-DD) [fact] text
-  Team memory uses ONLY the [fact] marker. Never use [pref] or [instr].
-- If a new fact contradicts an existing entry, update the existing entry.
-- Do not duplicate existing information.
-- Preserve neutral team phrasing; avoid person-specific memory unless role-anchored.
-
-If nothing is worth remembering, output exactly: NO_UPDATE
-
-<current_team_memory>
-{current_memory}
-</current_team_memory>
-
-<latest_message_author>
-{author}
-</latest_message_author>
-
-<latest_message>
-{user_message}
-</latest_message>"""
-
-
-async def extract_and_save_memory(
-    *,
-    user_message: str,
-    user_id: str | None,
-    llm: Any,
-) -> None:
-    """Background task: extract memorizable info and persist it.
-
-    Designed to be fire-and-forget — catches all exceptions internally.
-    """
-    if not user_id:
-        return
-
-    try:
-        uid = UUID(user_id) if isinstance(user_id, str) else user_id
-
-        async with shielded_async_session() as session:
-            result = await session.execute(select(User).where(User.id == uid))
-            user = result.scalars().first()
-            if not user:
-                return
-
-            old_memory = user.memory_md
-            first_name = (
-                user.display_name.strip().split()[0]
-                if user.display_name and user.display_name.strip()
-                else "The user"
-            )
-            prompt = _MEMORY_EXTRACT_PROMPT.format(
-                current_memory=old_memory or "(empty)",
-                user_message=user_message,
-                user_name=first_name,
-            )
-            response = await llm.ainvoke(
-                [HumanMessage(content=prompt)],
-                config={"tags": ["surfsense:internal", "memory-extraction"]},
-            )
-            text = extract_text_content(response.content).strip()
-
-            if text == "NO_UPDATE" or not text:
-                logger.debug("Memory extraction: no update needed (user %s)", uid)
-                return
-
-            save_result = await _save_memory(
-                updated_memory=text,
-                old_memory=old_memory,
-                llm=llm,
-                apply_fn=lambda content: setattr(user, "memory_md", content),
-                commit_fn=session.commit,
-                rollback_fn=session.rollback,
-                label="memory",
-                scope="user",
-            )
-            logger.info(
-                "Background memory extraction for user %s: %s",
-                uid,
-                save_result.get("status"),
-            )
-    except Exception:
-        logger.exception("Background user memory extraction failed")
-
-
-async def extract_and_save_team_memory(
-    *,
-    user_message: str,
-    search_space_id: int | None,
-    llm: Any,
-    author_display_name: str | None = None,
-) -> None:
-    """Background task: extract team-level memory and persist it.
-
-    Runs only for shared threads. Designed to be fire-and-forget and catches
-    exceptions internally.
-    """
-    if not search_space_id:
-        return
-
-    try:
-        async with shielded_async_session() as session:
-            result = await session.execute(
-                select(SearchSpace).where(SearchSpace.id == search_space_id)
-            )
-            space = result.scalars().first()
-            if not space:
-                return
-
-            old_memory = space.shared_memory_md
-            prompt = _TEAM_MEMORY_EXTRACT_PROMPT.format(
-                current_memory=old_memory or "(empty)",
-                author=author_display_name or "Unknown team member",
-                user_message=user_message,
-            )
-            response = await llm.ainvoke(
-                [HumanMessage(content=prompt)],
-                config={"tags": ["surfsense:internal", "team-memory-extraction"]},
-            )
-            text = extract_text_content(response.content).strip()
-
-            if text == "NO_UPDATE" or not text:
-                logger.debug(
-                    "Team memory extraction: no update needed (space %s)",
-                    search_space_id,
-                )
-                return
-
-            save_result = await _save_memory(
-                updated_memory=text,
-                old_memory=old_memory,
-                llm=llm,
-                apply_fn=lambda content: setattr(space, "shared_memory_md", content),
-                commit_fn=session.commit,
-                rollback_fn=session.rollback,
-                label="team memory",
-                scope="team",
-            )
-            logger.info(
-                "Background team memory extraction for space %s: %s",
-                search_space_id,
-                save_result.get("status"),
-            )
-    except Exception:
-        logger.exception("Background team memory extraction failed")

surfsense_backend/app/agents/new_chat/middleware/kb_persistence.py

@@ -32,6 +32,7 @@ exact same routine when ``aafter_agent`` was skipped (e.g. client disconnect).
 
 from __future__ import annotations
 
+import asyncio
 import logging
 from datetime import UTC, datetime
 from typing import Any
@@ -249,11 +250,11 @@ async def _create_document(
     session.add(doc)
     await session.flush()
 
-    summary_embedding = embed_texts([content])[0]
+    summary_embedding = (await asyncio.to_thread(embed_texts, [content]))[0]
     doc.embedding = summary_embedding
     chunks = chunk_text(content)
     if chunks:
-        chunk_embeddings = embed_texts(chunks)
+        chunk_embeddings = await asyncio.to_thread(embed_texts, chunks)
         session.add_all(
             [
                 Chunk(document_id=doc.id, content=text, embedding=embedding)
@@ -295,13 +296,13 @@ async def _update_document(
         search_space_id,
     )
 
-    summary_embedding = embed_texts([content])[0]
+    summary_embedding = (await asyncio.to_thread(embed_texts, [content]))[0]
     document.embedding = summary_embedding
 
     await session.execute(delete(Chunk).where(Chunk.document_id == document.id))
     chunks = chunk_text(content)
     if chunks:
-        chunk_embeddings = embed_texts(chunks)
+        chunk_embeddings = await asyncio.to_thread(embed_texts, chunks)
         session.add_all(
             [
                 Chunk(document_id=document.id, content=text, embedding=embedding)

surfsense_backend/app/agents/new_chat/middleware/knowledge_search.py

@@ -457,7 +457,7 @@ async def search_knowledge_base(
     if not query:
         return []
 
-    [embedding] = embed_texts([query])
+    [embedding] = await asyncio.to_thread(embed_texts, [query])
     doc_types = _resolve_search_types(available_connectors, available_document_types)
     retriever_top_k = min(top_k * 3, 30)
 
@@ -579,6 +579,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         self,
         *,
         llm: BaseChatModel | None = None,
+        planner_llm: BaseChatModel | None = None,
         search_space_id: int,
         filesystem_mode: FilesystemMode = FilesystemMode.CLOUD,
         available_connectors: list[str] | None = None,
@@ -588,6 +589,15 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         inject_system_message: bool = True,  # For backwards compatibility
     ) -> None:
         self.llm = llm
+        # The planner LLM handles short, structured internal tasks (query
+        # rewriting, date extraction, recency classification). When an
+        # operator marks a global config ``is_planner: true`` we route
+        # those calls to a cheap/fast model (e.g. gpt-4o-mini, Haiku, Azure
+        # gpt-5.x-nano) instead of the user's chat LLM — those classification
+        # tasks don't need frontier-tier capability. Falls back to the chat
+        # LLM when no planner config is wired up so deployments without one
+        # keep working unchanged.
+        self.planner_llm = planner_llm or llm
         self.search_space_id = search_space_id
         self.filesystem_mode = filesystem_mode
         self.available_connectors = available_connectors
@@ -598,7 +608,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         # Build the kb-planner private Runnable ONCE here so we don't pay
         # the ``create_agent`` compile cost (50-200ms) on every turn.
         # Disabled by default behind ``enable_kb_planner_runnable``; when
-        # off the planner falls back to the legacy ``self.llm.ainvoke``
+        # off the planner falls back to the legacy ``planner_llm.ainvoke``
         # path.
         self._planner: Runnable | None = None
         self._planner_compile_failed = False
@@ -608,7 +618,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
 
         Returns ``None`` when the feature flag is disabled, when the LLM is
         unavailable, or when ``create_agent`` raises (we fall back to the
-        legacy ``self.llm.ainvoke`` path in that case). Compilation happens
+        legacy ``planner_llm.ainvoke`` path in that case). Compilation happens
         lazily on first call, then memoized via ``self._planner``.
 
         The compiled agent is constructed without tools — the planner's
@@ -618,7 +628,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         """
         if self._planner is not None or self._planner_compile_failed:
             return self._planner
-        if self.llm is None:
+        if self.planner_llm is None:
             return None
         flags = get_flags()
         if not flags.enable_kb_planner_runnable or flags.disable_new_agent_stack:
@@ -628,13 +638,13 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
 
         try:
             self._planner = create_agent(
-                self.llm,
+                self.planner_llm,
                 tools=[],
                 middleware=[RetryAfterMiddleware(max_retries=2)],
             )
         except Exception as exc:  # pragma: no cover - defensive
             logger.warning(
-                "kb-planner Runnable compile failed; falling back to llm.ainvoke: %s",
+                "kb-planner Runnable compile failed; falling back to planner_llm.ainvoke: %s",
                 exc,
             )
             self._planner_compile_failed = True
@@ -647,12 +657,12 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         messages: Sequence[BaseMessage],
         user_text: str,
     ) -> tuple[str, datetime | None, datetime | None, bool]:
-        if self.llm is None:
+        if self.planner_llm is None:
             return user_text, None, None, False
 
         recent_conversation = _render_recent_conversation(
             messages,
-            llm=self.llm,
+            llm=self.planner_llm,
             user_text=user_text,
         )
         prompt = _build_kb_planner_prompt(
@@ -663,8 +673,8 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         t0 = loop.time()
 
         # Prefer the compiled-once planner Runnable when enabled; otherwise
-        # fall back to ``self.llm.ainvoke``. The ``surfsense:internal`` tag
-        # is preserved on both paths so ``_stream_agent_events`` still
+        # fall back to ``planner_llm.ainvoke``. The ``surfsense:internal``
+        # tag is preserved on both paths so ``_stream_agent_events`` still
         # suppresses the planner's intermediate events from the UI.
         planner = self._build_kb_planner_runnable()
         try:
@@ -684,7 +694,7 @@ class KnowledgePriorityMiddleware(AgentMiddleware):  # type: ignore[type-arg]
                     else AIMessage(content="")
                 )
             else:
-                response = await self.llm.ainvoke(
+                response = await self.planner_llm.ainvoke(
                     [HumanMessage(content=prompt)],
                     config={"tags": ["surfsense:internal"]},
                 )

surfsense_backend/app/agents/new_chat/middleware/knowledge_tree.py

@@ -24,6 +24,7 @@ from __future__ import annotations
 
 import asyncio
 import logging
+import time
 from typing import Any
 
 from langchain.agents.middleware import AgentMiddleware, AgentState
@@ -41,6 +42,9 @@ from app.agents.new_chat.path_resolver import (
     doc_to_virtual_path,
 )
 from app.db import Document, shielded_async_session
+from app.utils.perf import get_perf_logger
+
+_perf_log = get_perf_logger()
 
 try:
     from litellm import token_counter
@@ -124,6 +128,7 @@ class KnowledgeTreeMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         if self.filesystem_mode != FilesystemMode.CLOUD:
             return None
 
+        start = time.perf_counter()
         update: dict[str, Any] = {}
         if not state.get("cwd"):
             update["cwd"] = DOCUMENTS_ROOT
@@ -131,7 +136,11 @@ class KnowledgeTreeMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         anon_doc = state.get("kb_anon_doc")
         if anon_doc:
             tree_msg = self._render_anon_tree(anon_doc)
+            cache_outcome = "anon"
         else:
+            version = int(state.get("tree_version") or 0)
+            cache_key = (self.search_space_id, version, False)
+            cache_outcome = "hit" if cache_key in self._cache else "miss"
             tree_msg = await self._render_kb_tree(state)
 
         update["workspace_tree_text"] = tree_msg
@@ -141,6 +150,14 @@ class KnowledgeTreeMiddleware(AgentMiddleware):  # type: ignore[type-arg]
             insert_at = max(len(messages) - 1, 0)
             messages.insert(insert_at, SystemMessage(content=tree_msg))
             update["messages"] = messages
+
+        _perf_log.info(
+            "[knowledge_tree] cache=%s chars=%d elapsed=%.3fs space=%d",
+            cache_outcome,
+            len(tree_msg),
+            time.perf_counter() - start,
+            self.search_space_id,
+        )
         return update
 
     def before_agent(  # type: ignore[override]

surfsense_backend/app/agents/new_chat/middleware/memory_injection.py

@@ -8,6 +8,7 @@ Injects memory markdown into the system prompt on every turn:
 from __future__ import annotations
 
 import logging
+import time
 from typing import Any
 from uuid import UUID
 
@@ -17,10 +18,12 @@ from langgraph.runtime import Runtime
 from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
 
-from app.agents.new_chat.tools.update_memory import MEMORY_HARD_LIMIT, MEMORY_SOFT_LIMIT
 from app.db import ChatVisibility, SearchSpace, User, shielded_async_session
+from app.services.memory import MEMORY_HARD_LIMIT, MEMORY_SOFT_LIMIT
+from app.utils.perf import get_perf_logger
 
 logger = logging.getLogger(__name__)
+_perf_log = get_perf_logger()
 
 
 class MemoryInjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
@@ -53,9 +56,13 @@ class MemoryInjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         if not isinstance(last_message, HumanMessage):
             return None
 
+        start = time.perf_counter()
+        db_elapsed = 0.0
         memory_blocks: list[str] = []
+        scope = "team" if self.visibility == ChatVisibility.SEARCH_SPACE else "user"
 
         async with shielded_async_session() as session:
+            db_start = time.perf_counter()
             if self.visibility == ChatVisibility.SEARCH_SPACE:
                 team_memory = await self._load_team_memory(session)
                 if team_memory:
@@ -96,7 +103,15 @@ class MemoryInjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
                             f"</memory_warning>"
                         )
 
+        db_elapsed = time.perf_counter() - db_start
+
         if not memory_blocks:
+            _perf_log.info(
+                "[memory_injection] scope=%s injected=0 db=%.3fs total=%.3fs",
+                scope,
+                db_elapsed,
+                time.perf_counter() - start,
+            )
             return None
 
         memory_text = "\n\n".join(memory_blocks)
@@ -106,6 +121,13 @@ class MemoryInjectionMiddleware(AgentMiddleware):  # type: ignore[type-arg]
         insert_idx = 1 if len(new_messages) > 1 else 0
         new_messages.insert(insert_idx, memory_msg)
 
+        _perf_log.info(
+            "[memory_injection] scope=%s injected=1 chars=%d db=%.3fs total=%.3fs",
+            scope,
+            len(memory_text),
+            db_elapsed,
+            time.perf_counter() - start,
+        )
         return {"messages": new_messages}
 
     async def _load_user_memory(

surfsense_backend/app/agents/new_chat/prompt_caching.py

@@ -39,9 +39,19 @@ For OpenAI-family configs we additionally pass:
 
 - ``prompt_cache_key=f"surfsense-thread-{thread_id}"`` — routing hint that
   raises hit rate by sending requests with a shared prefix to the same
-  backend.
+  backend. Supported by ``openai/``, ``deepseek/``, ``xai/``, and
+  ``azure/`` (added to LiteLLM's Azure transformer in
+  https://github.com/BerriAI/litellm/pull/20989, Feb 2026; verified
+  against ``AzureOpenAIConfig.get_supported_openai_params`` in our
+  installed litellm 1.83.14 for ``azure/gpt-4o``, ``azure/gpt-4o-mini``,
+  ``azure/gpt-5.4``, ``azure/gpt-5.4-mini``).
 - ``prompt_cache_retention="24h"`` — extends cache TTL beyond the default
-  5-10 min in-memory cache.
+  5-10 min in-memory cache. Set ONLY for OpenAI/DeepSeek/xAI: Azure's
+  server-side support landed in Microsoft's docs on 2026-05-13 but
+  LiteLLM 1.83.14's Azure transformer still omits it from its supported
+  params list, so it gets silently dropped by ``litellm.drop_params``.
+  Azure's default in-memory retention (5-10 min, max 1 h) already
+  bridges intra-conversation turns; revisit when LiteLLM bumps Azure.
 
 Safety net: ``litellm.drop_params=True`` is set globally in
 ``app.services.llm_service`` at module-load time. Any kwarg the destination
@@ -81,13 +91,31 @@ _DEFAULT_INJECTION_POINTS: tuple[dict[str, Any], ...] = (
     {"location": "message", "index": -1},
 )
 
-# Providers (uppercase ``AgentConfig.provider`` values) that natively expose
-# OpenAI-style automatic prompt caching with ``prompt_cache_key`` and
-# ``prompt_cache_retention`` kwargs. Strict whitelist — many other providers
-# in ``PROVIDER_MAP`` route through litellm's ``openai`` prefix without
-# implementing the OpenAI prompt-cache surface (e.g. MOONSHOT, ZHIPU,
-# MINIMAX), so we can't infer family from the litellm prefix alone.
-_OPENAI_FAMILY_PROVIDERS: frozenset[str] = frozenset({"OPENAI", "DEEPSEEK", "XAI"})
+# Providers (uppercase ``AgentConfig.provider`` values) that accept the
+# OpenAI ``prompt_cache_key`` routing hint. Microsoft's Azure OpenAI docs
+# (2026-05-13) confirm automatic prompt caching applies to every GPT-4o
+# or newer Azure deployment at ≥1024 tokens with no configuration needed,
+# and that ``prompt_cache_key`` is combined with the prefix hash to
+# improve routing affinity and therefore cache hit rate. LiteLLM's Azure
+# transformer ships ``prompt_cache_key`` in its supported params as of
+# https://github.com/BerriAI/litellm/pull/20989.
+#
+# Strict whitelist — many other providers in ``PROVIDER_MAP`` route
+# through litellm's ``openai`` prefix without implementing the OpenAI
+# prompt-cache surface (e.g. MOONSHOT, ZHIPU, MINIMAX), so we can't infer
+# family from the litellm prefix alone.
+_PROMPT_CACHE_KEY_PROVIDERS: frozenset[str] = frozenset(
+    {"OPENAI", "DEEPSEEK", "XAI", "AZURE", "AZURE_OPENAI"}
+)
+
+# Subset of ``_PROMPT_CACHE_KEY_PROVIDERS`` that also accept
+# ``prompt_cache_retention="24h"``. Azure is excluded: see module
+# docstring — LiteLLM 1.83.14's Azure transformer omits the param so
+# ``drop_params`` silently strips it. Re-add Azure once a future LiteLLM
+# release wires it into ``AzureOpenAIConfig.get_supported_openai_params``.
+_PROMPT_CACHE_RETENTION_PROVIDERS: frozenset[str] = frozenset(
+    {"OPENAI", "DEEPSEEK", "XAI"}
+)
 
 
 def _is_router_llm(llm: BaseChatModel) -> bool:
@@ -101,13 +129,13 @@ def _is_router_llm(llm: BaseChatModel) -> bool:
     return type(llm).__name__ == "ChatLiteLLMRouter"
 
 
-def _is_openai_family_config(agent_config: AgentConfig | None) -> bool:
-    """Whether the config targets an OpenAI-style prompt-cache surface.
+def _provider_supports_prompt_cache_key(agent_config: AgentConfig | None) -> bool:
+    """Whether the config targets a provider that accepts ``prompt_cache_key``.
 
-    Strict — only returns True when the user explicitly chose OPENAI,
-    DEEPSEEK, or XAI as the provider in their ``NewLLMConfig`` /
-    ``YAMLConfig``. Auto-mode and custom providers return False because
-    we can't statically know the destination.
+    Strict — only returns True for explicitly chosen OPENAI, DEEPSEEK,
+    XAI, AZURE, or AZURE_OPENAI providers. Auto-mode and custom
+    providers return False because we can't statically know the
+    destination and the router fans out across mixed providers.
     """
     if agent_config is None or not agent_config.provider:
         return False
@@ -115,7 +143,25 @@ def _is_openai_family_config(agent_config: AgentConfig | None) -> bool:
         return False
     if agent_config.custom_provider:
         return False
-    return agent_config.provider.upper() in _OPENAI_FAMILY_PROVIDERS
+    return agent_config.provider.upper() in _PROMPT_CACHE_KEY_PROVIDERS
+
+
+def _provider_supports_prompt_cache_retention(
+    agent_config: AgentConfig | None,
+) -> bool:
+    """Whether the config targets a provider that accepts ``prompt_cache_retention``.
+
+    Tighter than :func:`_provider_supports_prompt_cache_key` — Azure
+    deployments are excluded until LiteLLM ships the param in its Azure
+    transformer (see module docstring).
+    """
+    if agent_config is None or not agent_config.provider:
+        return False
+    if agent_config.is_auto_mode:
+        return False
+    if agent_config.custom_provider:
+        return False
+    return agent_config.provider.upper() in _PROMPT_CACHE_RETENTION_PROVIDERS
 
 
 def _get_or_init_model_kwargs(llm: BaseChatModel) -> dict[str, Any] | None:
@@ -173,16 +219,23 @@ def apply_litellm_prompt_caching(
             dict(point) for point in _DEFAULT_INJECTION_POINTS
         ]
 
-    # OpenAI-family extras only when we statically know the destination is
-    # OpenAI / DeepSeek / xAI. Auto-mode router fans out across providers
-    # so we can't safely set OpenAI-only kwargs there (drop_params would
-    # strip them but it's wasteful to set them in the first place).
+    # OpenAI-style extras only when we statically know the destination
+    # accepts them. Auto-mode router fans out across mixed providers so
+    # we can't safely set destination-specific kwargs there (drop_params
+    # would strip them but it's wasteful to set them in the first
+    # place).
     if _is_router_llm(llm):
         return
-    if not _is_openai_family_config(agent_config):
-        return
 
-    if thread_id is not None and "prompt_cache_key" not in model_kwargs:
+    if (
+        thread_id is not None
+        and "prompt_cache_key" not in model_kwargs
+        and _provider_supports_prompt_cache_key(agent_config)
+    ):
         model_kwargs["prompt_cache_key"] = f"surfsense-thread-{thread_id}"
-    if "prompt_cache_retention" not in model_kwargs:
+
+    if (
+        "prompt_cache_retention" not in model_kwargs
+        and _provider_supports_prompt_cache_retention(agent_config)
+    ):
         model_kwargs["prompt_cache_retention"] = "24h"

surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_private.md

@@ -3,4 +3,10 @@ IMPORTANT — After understanding each user message, ALWAYS check: does this mes
 reveal durable facts about the user (role, interests, preferences, projects,
 background, or standing instructions)? If yes, you MUST call update_memory
 alongside your normal response — do not defer this to a later turn.
+
+Memory is stored as a heading-based markdown document. New entries should be
+under `##` headings such as `## Facts`, `## Preferences`, or `## Instructions`
+with bullets like `- YYYY-MM-DD: text`. If existing memory contains legacy
+`(YYYY-MM-DD) [fact|pref|instr]` markers, preserve the information but write
+new saves in the heading-based format.
 </memory_protocol>

surfsense_backend/app/agents/new_chat/prompts/base/memory_protocol_team.md

@@ -3,4 +3,12 @@ IMPORTANT — After understanding each user message, ALWAYS check: does this mes
 reveal durable facts about the team (decisions, conventions, architecture, processes,
 or key facts)? If yes, you MUST call update_memory alongside your normal response —
 do not defer this to a later turn.
+
+Team memory is stored as a heading-based markdown document. New entries should
+be under `##` headings such as `## Product Decisions`,
+`## Engineering Conventions`, `## Project Facts`, or `## Open Questions` with
+bullets like `- YYYY-MM-DD: text`. If existing memory contains legacy
+`(YYYY-MM-DD) [fact]` markers, preserve the information but write new saves in
+the heading-based format. Do not create personal headings such as
+`## Preferences` or `## Instructions`.
 </memory_protocol>

surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_private.md

@@ -1,16 +1,16 @@
 
 - <user_name>Alex</user_name>, <user_memory> is empty. User: "I'm a space enthusiast, explain astrophage to me"
-  - The user casually shared a durable fact. Use their first name in the entry, short neutral heading:
-    update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n")
+  - The user casually shared a durable fact:
+    update_memory(updated_memory="## Facts\n- 2025-03-15: Alex is a space enthusiast\n")
 - User: "Remember that I prefer concise answers over detailed explanations"
-  - Durable preference. Merge with existing memory, add a new heading:
-    update_memory(updated_memory="## Interests & background\n- (2025-03-15) [fact] Alex is a space enthusiast\n\n## Response style\n- (2025-03-15) [pref] Alex prefers concise answers over detailed explanations\n")
+  - Durable preference. Merge with existing memory:
+    update_memory(updated_memory="## Facts\n- 2025-03-15: Alex is a space enthusiast\n\n## Preferences\n- 2025-03-15: Alex prefers concise answers over detailed explanations\n")
 - User: "I actually moved to Tokyo last month"
   - Updated fact, date prefix reflects when recorded:
-    update_memory(updated_memory="## Interests & background\n...\n\n## Personal context\n- (2025-03-15) [fact] Alex lives in Tokyo (previously London)\n...")
+    update_memory(updated_memory="## Facts\n- 2025-03-15: Alex lives in Tokyo (previously London)\n...")
 - User: "I'm a freelance photographer working on a nature documentary"
   - Durable background info under a fitting heading:
-    update_memory(updated_memory="...\n\n## Current focus\n- (2025-03-15) [fact] Alex is a freelance photographer\n- (2025-03-15) [fact] Alex is working on a nature documentary\n")
+    update_memory(updated_memory="...\n\n## Current Focus\n- 2025-03-15: Alex is a freelance photographer\n- 2025-03-15: Alex is working on a nature documentary\n")
 - User: "Always respond in bullet points"
   - Standing instruction:
-    update_memory(updated_memory="...\n\n## Response style\n- (2025-03-15) [instr] Always respond to Alex in bullet points\n")
+    update_memory(updated_memory="...\n\n## Instructions\n- 2025-03-15: Always respond to Alex in bullet points\n")

surfsense_backend/app/agents/new_chat/prompts/examples/update_memory_team.md

@@ -1,7 +1,7 @@
 
 - User: "Let's remember that we decided to do weekly standup meetings on Mondays"
   - Durable team decision:
-    update_memory(updated_memory="- (2025-03-15) [fact] Weekly standup meetings on Mondays\n...")
+    update_memory(updated_memory="## Product Decisions\n- 2025-03-15: Weekly standup meetings happen on Mondays\n...")
 - User: "Our office is in downtown Seattle, 5th floor"
   - Durable team fact:
-    update_memory(updated_memory="- (2025-03-15) [fact] Office location: downtown Seattle, 5th floor\n...")
+    update_memory(updated_memory="## Project Facts\n- 2025-03-15: Office location is downtown Seattle, 5th floor\n...")

surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_private.md

@@ -1,31 +1,26 @@
 
 - update_memory: Update your personal memory document about the user.
-  - Your current memory is already in <user_memory> in your context.  The `chars` and
-    `limit` attributes show your current usage and the maximum allowed size.
-  - This is your curated long-term memory — the distilled essence of what you know about
-    the user, not raw conversation logs.
-  - Call update_memory when:
-    * The user explicitly asks to remember or forget something
-    * The user shares durable facts or preferences that will matter in future conversations
-  - The user's first name is provided in <user_name>. Use it in memory entries
-    instead of "the user" (e.g. "{name} works at..." not "The user works at...").
-    Do not store the name itself as a separate memory entry.
-  - Do not store short-lived or ephemeral info: one-off questions, greetings,
-    session logistics, or things that only matter for the current task.
+  - Your current memory is already in <user_memory> in your context. The `chars`
+    and `limit` attributes show current usage and the maximum allowed size.
+  - This is curated long-term memory, not raw conversation logs.
+  - Call update_memory when the user explicitly asks to remember/forget
+    something or shares durable facts, preferences, or standing instructions.
+  - The user's first name is provided in <user_name>. Use it in entries instead
+    of "the user" when helpful. Do not store the name alone as a memory entry.
+  - Do not store short-lived info: one-off questions, greetings, session
+    logistics, or things that only matter for the current task.
   - Args:
-    - updated_memory: The FULL updated markdown document (not a diff).
-      Merge new facts with existing ones, update contradictions, remove outdated entries.
-      Treat every update as a curation pass — consolidate, don't just append.
-  - Every bullet MUST use this format: - (YYYY-MM-DD) [marker] text
-    Markers:
-      [fact]  — durable facts (role, background, projects, tools, expertise)
-      [pref]  — preferences (response style, languages, formats, tools)
-      [instr] — standing instructions (always/never do, response rules)
-  - Keep it concise and well under the character limit shown in <user_memory>.
-  - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and
-    natural. Do NOT include the user's name in headings. Organize by context — e.g.
-    who they are, what they're focused on, how they prefer things. Create, split, or
-    merge headings freely as the memory grows.
-  - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant
-    details and context rather than just a few words.
-  - During consolidation, prioritize keeping: [instr] > [pref] > [fact].
+    - updated_memory: The FULL updated markdown document, not a diff. Merge new
+      facts with existing ones, update contradictions, remove outdated entries,
+      and consolidate instead of only appending.
+  - Use heading-based Markdown:
+    * Every entry must be under a `##` heading.
+    * Recommended headings: `## Facts`, `## Preferences`, `## Instructions`.
+      Specific natural headings are allowed when clearer.
+    * New bullets should use `- YYYY-MM-DD: text`.
+    * Each entry should be one concise but descriptive bullet.
+  - If existing memory uses legacy `(YYYY-MM-DD) [fact|pref|instr]` markers,
+    preserve the information but write the updated document in the new
+    heading-based format.
+  - During consolidation, prioritize durable instructions and preferences before
+    generic facts.

surfsense_backend/app/agents/new_chat/prompts/tools/update_memory_team.md

@@ -1,26 +1,28 @@
 
 - update_memory: Update the team's shared memory document for this search space.
-  - Your current team memory is already in <team_memory> in your context.  The `chars`
-    and `limit` attributes show current usage and the maximum allowed size.
-  - This is the team's curated long-term memory — decisions, conventions, key facts.
-  - NEVER store personal memory in team memory (e.g. personal bio, individual
-    preferences, or user-only standing instructions).
-  - Call update_memory when:
-    * A team member explicitly asks to remember or forget something
-    * The conversation surfaces durable team decisions, conventions, or facts
-      that will matter in future conversations
-  - Do not store short-lived or ephemeral info: one-off questions, greetings,
-    session logistics, or things that only matter for the current task.
+  - Your current team memory is already in <team_memory> in your context. The
+    `chars` and `limit` attributes show current usage and the maximum allowed size.
+  - This is curated long-term team memory: decisions, conventions, architecture,
+    processes, and key shared facts.
+  - NEVER store personal memory in team memory: individual bios, personal
+    preferences, or user-only standing instructions.
+  - Call update_memory when a team member asks to remember/forget something, or
+    when the conversation surfaces durable team context that matters later.
+  - Do not store short-lived info: one-off questions, greetings, session
+    logistics, or things that only matter for the current task.
   - Args:
-    - updated_memory: The FULL updated markdown document (not a diff).
-      Merge new facts with existing ones, update contradictions, remove outdated entries.
-      Treat every update as a curation pass — consolidate, don't just append.
-  - Every bullet MUST use this format: - (YYYY-MM-DD) [fact] text
-    Team memory uses ONLY the [fact] marker. Never use [pref] or [instr] in team memory.
-  - Keep it concise and well under the character limit shown in <team_memory>.
-  - Every entry MUST be under a `##` heading. Keep heading names short (2-3 words) and
-    natural. Organize by context — e.g. what the team decided, current architecture,
-    active processes. Create, split, or merge headings freely as the memory grows.
-  - Each entry MUST be a single bullet point. Be descriptive but concise — include relevant
-    details and context rather than just a few words.
-  - During consolidation, prioritize keeping: decisions/conventions > key facts > current priorities.
+    - updated_memory: The FULL updated markdown document, not a diff. Merge new
+      facts with existing ones, update contradictions, remove outdated entries,
+      and consolidate instead of only appending.
+  - Use heading-based Markdown:
+    * Every entry must be under a `##` heading.
+    * Recommended headings: `## Product Decisions`, `## Engineering Conventions`,
+      `## Project Facts`, `## Open Questions`.
+    * New bullets should use `- YYYY-MM-DD: text`.
+    * Each entry should be one concise but descriptive bullet.
+  - If existing memory uses legacy `(YYYY-MM-DD) [fact]` markers, preserve the
+    information but write the updated document in the new heading-based format.
+  - Do not create personal headings such as `## Preferences`, `## Instructions`,
+    `## Personal Notes`, or `## Personal Instructions`.
+  - During consolidation, prioritize decisions/conventions, then key facts, then
+    current priorities.

surfsense_backend/app/agents/new_chat/tools/mcp_tool.py

@@ -36,8 +36,16 @@ from sqlalchemy.ext.asyncio import AsyncSession
 from app.agents.new_chat.middleware.dedup_tool_calls import dedup_key_full_args
 from app.agents.new_chat.tools.hitl import request_approval
 from app.agents.new_chat.tools.mcp_client import MCPClient
+from app.agents.new_chat.tools.mcp_tools_cache import (
+    CachedMCPTools,
+    read_cached_tools,
+    write_cached_tools,
+)
 from app.db import SearchSourceConnector
 from app.services.mcp_oauth.registry import MCP_SERVICES, get_service_by_connector_type
+from app.utils.perf import get_perf_logger
+
+_perf_log = get_perf_logger()
 
 logger = logging.getLogger(__name__)
 
@@ -293,15 +301,21 @@ async def _create_mcp_tool_from_definition_http(
         timeout: float = 60.0,
     ) -> str:
         """Execute a single MCP HTTP call with the given headers."""
+        call_start = time.perf_counter()
         async with (
             streamablehttp_client(url, headers=call_headers) as (read, write, _),
             ClientSession(read, write) as session,
         ):
+            init_start = time.perf_counter()
             await session.initialize()
+            init_elapsed = time.perf_counter() - init_start
+
+            tool_start = time.perf_counter()
             response = await asyncio.wait_for(
                 session.call_tool(original_tool_name, arguments=call_kwargs),
                 timeout=timeout,
             )
+            tool_elapsed = time.perf_counter() - tool_start
 
             result = []
             for content in response.content:
@@ -312,7 +326,18 @@ async def _create_mcp_tool_from_definition_http(
                 else:
                     result.append(str(content))
 
-            return "\n".join(result) if result else ""
+            payload = "\n".join(result) if result else ""
+
+        _perf_log.info(
+            "[mcp_http_call] connector=%s tool=%s init=%.3fs call=%.3fs total=%.3fs out_chars=%d",
+            connector_id,
+            original_tool_name,
+            init_elapsed,
+            tool_elapsed,
+            time.perf_counter() - call_start,
+            len(payload),
+        )
+        return payload
 
     async def mcp_http_tool_call(**kwargs) -> str:
         """Execute the MCP tool call via HTTP transport."""
@@ -496,6 +521,7 @@ async def _load_http_mcp_tools(
     is_generic_mcp: bool = False,
     *,
     bypass_internal_hitl: bool = False,
+    cached_tools: CachedMCPTools | None = None,
 ) -> list[StructuredTool]:
     """Load tools from an HTTP-based MCP server.
 
@@ -506,6 +532,8 @@ async def _load_http_mcp_tools(
         readonly_tools: Tool names that skip HITL approval (read-only operations).
         tool_name_prefix: If set, each tool name is prefixed for multi-account
             disambiguation (e.g. ``linear_25``).
+        cached_tools: If provided, skip live discovery and rebuild wrappers
+            from the persisted definitions.
     """
     tools: list[StructuredTool] = []
 
@@ -529,15 +557,23 @@ async def _load_http_mcp_tools(
 
     allowed_set = set(allowed_tools) if allowed_tools else None
 
-    async def _discover(disc_headers: dict[str, str]) -> list[dict[str, Any]]:
-        """Connect, initialize, and list tools from the MCP server."""
+    async def _discover(
+        disc_headers: dict[str, str],
+    ) -> tuple[dict[str, str | None], list[dict[str, Any]]]:
+        """Connect, initialize, and list tools — returns (serverInfo, tools)."""
         async with (
             streamablehttp_client(url, headers=disc_headers) as (read, write, _),
             ClientSession(read, write) as session,
         ):
-            await session.initialize()
+            init_result = await session.initialize()
+            server_info: dict[str, str | None] = {"name": None, "version": None}
+            si = getattr(init_result, "serverInfo", None)
+            if si is not None:
+                server_info["name"] = getattr(si, "name", None)
+                server_info["version"] = getattr(si, "version", None)
+
             response = await session.list_tools()
-            return [
+            return server_info, [
                 {
                     "name": tool.name,
                     "description": tool.description or "",
@@ -548,47 +584,65 @@ async def _load_http_mcp_tools(
                 for tool in response.tools
             ]
 
-    try:
-        tool_definitions = await _discover(headers)
-    except Exception as first_err:
-        if not _is_auth_error(first_err) or connector_id is None:
-            logger.exception(
-                "Failed to connect to HTTP MCP server at '%s' (connector %d): %s",
-                url,
-                connector_id,
-                first_err,
-            )
-            return tools
-
-        logger.warning(
-            "HTTP MCP discovery for connector %d got 401 — attempting token refresh",
-            connector_id,
-        )
-        fresh_headers = await _force_refresh_and_get_headers(connector_id)
-        if fresh_headers is None:
-            await _mark_connector_auth_expired(connector_id)
-            logger.error(
-                "HTTP MCP discovery for connector %d: token refresh failed, marking auth_expired",
-                connector_id,
-            )
-            return tools
-
+    if cached_tools is not None:
+        tool_definitions = [
+            {
+                "name": td.name,
+                "description": td.description,
+                "input_schema": td.input_schema,
+            }
+            for td in cached_tools.tools
+        ]
+    else:
         try:
-            tool_definitions = await _discover(fresh_headers)
-            headers = fresh_headers
-            logger.info(
-                "HTTP MCP discovery for connector %d succeeded after 401 recovery",
+            server_info, tool_definitions = await _discover(headers)
+        except Exception as first_err:
+            if not _is_auth_error(first_err) or connector_id is None:
+                logger.exception(
+                    "Failed to connect to HTTP MCP server at '%s' (connector %d): %s",
+                    url,
+                    connector_id,
+                    first_err,
+                )
+                return tools
+
+            logger.warning(
+                "HTTP MCP discovery for connector %d got 401 — attempting token refresh",
                 connector_id,
             )
-        except Exception as retry_err:
-            logger.exception(
-                "HTTP MCP discovery for connector %d still failing after refresh: %s",
-                connector_id,
-                retry_err,
-            )
-            if _is_auth_error(retry_err):
+            fresh_headers = await _force_refresh_and_get_headers(connector_id)
+            if fresh_headers is None:
                 await _mark_connector_auth_expired(connector_id)
-            return tools
+                logger.error(
+                    "HTTP MCP discovery for connector %d: token refresh failed, marking auth_expired",
+                    connector_id,
+                )
+                return tools
+
+            try:
+                server_info, tool_definitions = await _discover(fresh_headers)
+                headers = fresh_headers
+                logger.info(
+                    "HTTP MCP discovery for connector %d succeeded after 401 recovery",
+                    connector_id,
+                )
+            except Exception as retry_err:
+                logger.exception(
+                    "HTTP MCP discovery for connector %d still failing after refresh: %s",
+                    connector_id,
+                    retry_err,
+                )
+                if _is_auth_error(retry_err):
+                    await _mark_connector_auth_expired(connector_id)
+                return tools
+
+        await write_cached_tools(
+            connector_id,
+            tool_definitions,
+            server_name=server_info.get("name"),
+            server_version=server_info.get("version"),
+            transport=server_config.get("transport", "streamable-http"),
+        )
 
     total_discovered = len(tool_definitions)
 
@@ -792,14 +846,25 @@ async def _maybe_refresh_mcp_oauth_token(
     except (ValueError, TypeError):
         return server_config
 
+    refresh_start = time.perf_counter()
     try:
         new_access = await _refresh_connector_token(session, connector)
         if not new_access:
+            _perf_log.info(
+                "[mcp_oauth_refresh] connector=%s elapsed=%.3fs outcome=no_token",
+                connector.id,
+                time.perf_counter() - refresh_start,
+            )
             return server_config
 
         logger.info(
             "Proactively refreshed MCP OAuth token for connector %s", connector.id
         )
+        _perf_log.info(
+            "[mcp_oauth_refresh] connector=%s elapsed=%.3fs outcome=refreshed",
+            connector.id,
+            time.perf_counter() - refresh_start,
+        )
 
         refreshed_config = dict(server_config)
         refreshed_config["headers"] = {
@@ -809,6 +874,11 @@ async def _maybe_refresh_mcp_oauth_token(
         return refreshed_config
 
     except Exception:
+        _perf_log.info(
+            "[mcp_oauth_refresh] connector=%s elapsed=%.3fs outcome=failed",
+            connector.id,
+            time.perf_counter() - refresh_start,
+        )
         logger.warning(
             "Failed to refresh MCP OAuth token for connector %s",
             connector.id,
@@ -937,6 +1007,94 @@ def invalidate_mcp_tools_cache(search_space_id: int | None = None) -> None:
         _mcp_tools_cache.clear()
 
 
+async def discover_single_mcp_connector(connector_id: int) -> None:
+    """Force live MCP discovery for one connector so its ``cached_tools`` row is fresh.
+
+    ``_load_http_mcp_tools`` persists ``cached_tools`` as a side effect of any
+    live discovery; passing ``cached_tools=None`` here guarantees we go to the
+    network. The returned wrappers are discarded — the in-process LRU is
+    rebuilt lazily on the next user query. Stdio connectors are not cached and
+    are skipped.
+    """
+    from app.db import async_session_maker
+
+    started = time.perf_counter()
+    try:
+        async with async_session_maker() as session:
+            connector = await session.get(SearchSourceConnector, connector_id)
+            if connector is None:
+                logger.info(
+                    "discover_single_mcp_connector: connector %d not found",
+                    connector_id,
+                )
+                return
+
+            cfg = connector.config or {}
+            server_config = cfg.get("server_config", {})
+            if not server_config or not isinstance(server_config, dict):
+                return
+
+            transport = server_config.get("transport", "stdio")
+            if transport not in ("streamable-http", "http", "sse"):
+                return
+
+            if cfg.get("mcp_oauth"):
+                server_config = await _maybe_refresh_mcp_oauth_token(
+                    session, connector, cfg, server_config
+                )
+                cfg = connector.config or {}
+                server_config = _inject_oauth_headers(cfg, server_config)
+                if server_config is None:
+                    logger.info(
+                        "discover_single_mcp_connector: OAuth token unavailable for connector %d",
+                        connector_id,
+                    )
+                    return
+
+            ct = (
+                connector.connector_type.value
+                if hasattr(connector.connector_type, "value")
+                else str(connector.connector_type)
+            )
+            svc_cfg = get_service_by_connector_type(ct)
+            allowed_tools = svc_cfg.allowed_tools if svc_cfg else []
+            readonly_tools = svc_cfg.readonly_tools if svc_cfg else frozenset()
+
+            await asyncio.wait_for(
+                _load_http_mcp_tools(
+                    connector.id,
+                    connector.name,
+                    server_config,
+                    trusted_tools=cfg.get("trusted_tools", []),
+                    allowed_tools=allowed_tools,
+                    readonly_tools=readonly_tools,
+                    tool_name_prefix=None,
+                    is_generic_mcp=svc_cfg is None,
+                    bypass_internal_hitl=True,
+                    cached_tools=None,
+                ),
+                timeout=_MCP_DISCOVERY_TIMEOUT_SECONDS,
+            )
+
+            _perf_log.info(
+                "[mcp_prefetch] connector=%s elapsed=%.3fs",
+                connector_id,
+                time.perf_counter() - started,
+            )
+    except TimeoutError:
+        logger.warning(
+            "discover_single_mcp_connector: connector %d timed out after %ds",
+            connector_id,
+            _MCP_DISCOVERY_TIMEOUT_SECONDS,
+        )
+    except Exception:
+        logger.warning(
+            "discover_single_mcp_connector: failed for connector %d",
+            connector_id,
+            exc_info=True,
+        )
+
+
 async def load_mcp_tools(
     session: AsyncSession,
     search_space_id: int,
@@ -1063,6 +1221,7 @@ async def load_mcp_tools(
                         "tool_name_prefix": tool_name_prefix,
                         "transport": server_config.get("transport", "stdio"),
                         "is_generic_mcp": svc_cfg is None,
+                        "cached_tools": read_cached_tools(connector),
                     }
                 )
 
@@ -1074,9 +1233,12 @@ async def load_mcp_tools(
                 )
 
         async def _discover_one(task: dict[str, Any]) -> list[StructuredTool]:
+            discover_start = time.perf_counter()
+            transport = task["transport"]
+            cached_tools = task.get("cached_tools")
             try:
-                if task["transport"] in ("streamable-http", "http", "sse"):
-                    return await asyncio.wait_for(
+                if transport in ("streamable-http", "http", "sse"):
+                    result = await asyncio.wait_for(
                         _load_http_mcp_tools(
                             task["connector_id"],
                             task["connector_name"],
@@ -1087,11 +1249,12 @@ async def load_mcp_tools(
                             tool_name_prefix=task["tool_name_prefix"],
                             is_generic_mcp=task.get("is_generic_mcp", False),
                             bypass_internal_hitl=bypass_internal_hitl,
+                            cached_tools=cached_tools,
                         ),
                         timeout=_MCP_DISCOVERY_TIMEOUT_SECONDS,
                     )
                 else:
-                    return await asyncio.wait_for(
+                    result = await asyncio.wait_for(
                         _load_stdio_mcp_tools(
                             task["connector_id"],
                             task["connector_name"],
@@ -1101,7 +1264,24 @@ async def load_mcp_tools(
                         ),
                         timeout=_MCP_DISCOVERY_TIMEOUT_SECONDS,
                     )
+                _perf_log.info(
+                    "[mcp_discover] connector=%s name=%r transport=%s tools=%d elapsed=%.3fs cache=%s",
+                    task["connector_id"],
+                    task["connector_name"],
+                    transport,
+                    len(result),
+                    time.perf_counter() - discover_start,
+                    "hit" if cached_tools is not None else "miss",
+                )
+                return result
             except TimeoutError:
+                _perf_log.info(
+                    "[mcp_discover] connector=%s name=%r transport=%s elapsed=%.3fs outcome=timeout",
+                    task["connector_id"],
+                    task["connector_name"],
+                    transport,
+                    time.perf_counter() - discover_start,
+                )
                 logger.error(
                     "MCP connector %d timed out after %ds during discovery",
                     task["connector_id"],
@@ -1109,6 +1289,13 @@ async def load_mcp_tools(
                 )
                 return []
             except Exception as e:
+                _perf_log.info(
+                    "[mcp_discover] connector=%s name=%r transport=%s elapsed=%.3fs outcome=error",
+                    task["connector_id"],
+                    task["connector_name"],
+                    transport,
+                    time.perf_counter() - discover_start,
+                )
                 logger.exception(
                     "Failed to load tools from MCP connector %d: %s",
                     task["connector_id"],
@@ -1116,7 +1303,14 @@ async def load_mcp_tools(
                 )
                 return []
 
+        gather_start = time.perf_counter()
         results = await asyncio.gather(*[_discover_one(t) for t in discovery_tasks])
+        _perf_log.info(
+            "[mcp_discover] gather_wall=%.3fs connectors=%d total_tools=%d",
+            time.perf_counter() - gather_start,
+            len(discovery_tasks),
+            sum(len(r) for r in results),
+        )
         tools: list[StructuredTool] = [tool for sublist in results for tool in sublist]
 
         _mcp_tools_cache[cache_key] = (now, tools)

surfsense_backend/app/agents/new_chat/tools/mcp_tools_cache.py

@@ -0,0 +1,145 @@
+"""Persist MCP ``list_tools`` results in ``SearchSourceConnector.config.cached_tools``."""
+
+from __future__ import annotations
+
+import asyncio
+import logging
+from datetime import UTC, datetime
+from typing import Any
+
+from pydantic import BaseModel, Field, ValidationError
+from sqlalchemy import select
+from sqlalchemy.orm.attributes import flag_modified
+
+from app.db import SearchSourceConnector, async_session_maker
+
+logger = logging.getLogger(__name__)
+
+_pending_prefetch_tasks: set[asyncio.Task[None]] = set()
+
+
+class CachedMCPToolDef(BaseModel):
+    name: str
+    description: str = ""
+    input_schema: dict[str, Any] = Field(default_factory=dict)
+
+
+class CachedMCPTools(BaseModel):
+    discovered_at: datetime
+    server_version: str | None = None
+    server_name: str | None = None
+    transport: str | None = None
+    tools: list[CachedMCPToolDef]
+
+
+def read_cached_tools(connector: SearchSourceConnector) -> CachedMCPTools | None:
+    """Return parsed cached tools or ``None`` if missing / corrupt (caller falls back to live discovery)."""
+    cfg = connector.config or {}
+    raw = cfg.get("cached_tools")
+    if not raw or not isinstance(raw, dict):
+        return None
+
+    try:
+        return CachedMCPTools.model_validate(raw)
+    except ValidationError as exc:
+        logger.warning(
+            "MCP connector %d has corrupt cached_tools — falling back to live discovery: %s",
+            connector.id,
+            exc,
+        )
+        return None
+
+
+async def write_cached_tools(
+    connector_id: int,
+    tool_definitions: list[dict[str, Any]],
+    *,
+    server_name: str | None = None,
+    server_version: str | None = None,
+    transport: str | None = None,
+) -> None:
+    """Best-effort persist; uses its own session so a write failure cannot poison the caller's transaction."""
+    payload = CachedMCPTools(
+        discovered_at=datetime.now(UTC),
+        server_version=server_version,
+        server_name=server_name,
+        transport=transport,
+        tools=[CachedMCPToolDef.model_validate(td) for td in tool_definitions],
+    )
+
+    try:
+        async with async_session_maker() as session:
+            result = await session.execute(
+                select(SearchSourceConnector).filter(
+                    SearchSourceConnector.id == connector_id,
+                )
+            )
+            connector = result.scalars().first()
+            if connector is None:
+                return
+
+            cfg = dict(connector.config or {})
+            cfg["cached_tools"] = payload.model_dump(mode="json")
+            connector.config = cfg
+            flag_modified(connector, "config")
+            await session.commit()
+
+            logger.info(
+                "Persisted cached_tools for MCP connector %d (%d tools)",
+                connector_id,
+                len(payload.tools),
+            )
+    except Exception:
+        logger.warning(
+            "Failed to persist cached_tools for MCP connector %d",
+            connector_id,
+            exc_info=True,
+        )
+
+
+def refresh_mcp_tools_cache_for_connector(
+    connector_id: int,
+    search_space_id: int,
+) -> None:
+    """Maintain the MCP tool cache after a single-connector lifecycle event.
+
+    Synchronously evicts the in-process LRU for the connector's search space
+    (LRU keys are per-space, so eviction cannot be scoped finer), then schedules
+    a background live discovery for this connector alone so its persisted
+    ``cached_tools`` row is refreshed before the next user query.
+
+    Idempotent. Eviction is best-effort; prefetch is best-effort and only runs
+    when an event loop is available. Neither path raises.
+    """
+    try:
+        from app.agents.new_chat.tools.mcp_tool import invalidate_mcp_tools_cache
+
+        invalidate_mcp_tools_cache(search_space_id)
+    except Exception:
+        logger.debug(
+            "MCP in-process cache eviction skipped for space %d",
+            search_space_id,
+            exc_info=True,
+        )
+
+    try:
+        loop = asyncio.get_running_loop()
+    except RuntimeError:
+        return
+
+    task = loop.create_task(_run_connector_prefetch(connector_id))
+    _pending_prefetch_tasks.add(task)
+    task.add_done_callback(_pending_prefetch_tasks.discard)
+
+
+async def _run_connector_prefetch(connector_id: int) -> None:
+    from app.agents.new_chat.tools.mcp_tool import discover_single_mcp_connector
+
+    try:
+        await discover_single_mcp_connector(connector_id)
+    except Exception:
+        logger.warning(
+            "MCP background prefetch failed for connector_id=%d",
+            connector_id,
+            exc_info=True,
+        )

surfsense_backend/app/agents/new_chat/tools/update_memory.py

@@ -1,369 +1,53 @@
-"""Markdown-document memory tool for the SurfSense agent.
-
-Replaces the old row-per-fact save_memory / recall_memory tools with a single
-update_memory tool that overwrites a freeform markdown TEXT column.  The LLM
-always sees the current memory in <user_memory> / <team_memory> tags injected
-by MemoryInjectionMiddleware, so it passes the FULL updated document each time.
-
-Overflow handling:
-  - Soft limit (18K chars): a warning is returned telling the agent to
-    consolidate on the next update.
-  - Hard limit (25K chars): a forced LLM-driven rewrite compresses the document.
-    If it still exceeds the limit after rewriting, the save is rejected.
-  - Diff validation: warns when entire ``##`` sections are dropped or when the
-    document shrinks by more than 60%.
-"""
+"""Memory update tools backed by the canonical memory service."""
 
 from __future__ import annotations
 
 import logging
-import re
-from typing import Any, Literal
+from typing import Any
 from uuid import UUID
 
-from langchain_core.messages import HumanMessage
 from langchain_core.tools import tool
-from sqlalchemy import select
 from sqlalchemy.ext.asyncio import AsyncSession
 
-from app.db import SearchSpace, User, async_session_maker
-from app.utils.content_utils import extract_text_content
+from app.db import async_session_maker
+from app.services.memory import MemoryScope, save_memory
 
 logger = logging.getLogger(__name__)
 
-MEMORY_SOFT_LIMIT = 18_000
-MEMORY_HARD_LIMIT = 25_000
-
-_SECTION_HEADING_RE = re.compile(r"^##\s+(.+)$", re.MULTILINE)
-_HEADING_NORMALIZE_RE = re.compile(r"\s+")
-
-_MARKER_RE = re.compile(r"\[(fact|pref|instr)\]")
-_BULLET_FORMAT_RE = re.compile(r"^- \(\d{4}-\d{2}-\d{2}\) \[(fact|pref|instr)\] .+$")
-_PERSONAL_ONLY_MARKERS = {"pref", "instr"}
-
-
-# ---------------------------------------------------------------------------
-# Diff validation
-# ---------------------------------------------------------------------------
-
-
-def _extract_headings(memory: str) -> set[str]:
-    """Return all ``## …`` heading texts (without the ``## `` prefix)."""
-    return set(_SECTION_HEADING_RE.findall(memory))
-
-
-def _normalize_heading(heading: str) -> str:
-    """Normalize heading text for robust scope checks."""
-    return _HEADING_NORMALIZE_RE.sub(" ", heading.strip().lower())
-
-
-def _validate_memory_scope(
-    content: str, scope: Literal["user", "team"]
-) -> dict[str, Any] | None:
-    """Reject personal-only markers ([pref], [instr]) in team memory."""
-    if scope != "team":
-        return None
-
-    markers = set(_MARKER_RE.findall(content))
-    leaked = sorted(markers & _PERSONAL_ONLY_MARKERS)
-    if leaked:
-        tags = ", ".join(f"[{m}]" for m in leaked)
-        return {
-            "status": "error",
-            "message": (
-                f"Team memory cannot include personal markers: {tags}. "
-                "Use [fact] only in team memory."
-            ),
-        }
-    return None
-
-
-def _validate_bullet_format(content: str) -> list[str]:
-    """Return warnings for bullet lines that don't match the required format.
-
-    Expected: ``- (YYYY-MM-DD) [fact|pref|instr] text``
-    """
-    warnings: list[str] = []
-    for line in content.splitlines():
-        stripped = line.strip()
-        if not stripped.startswith("- "):
-            continue
-        if not _BULLET_FORMAT_RE.match(stripped):
-            short = stripped[:80] + ("..." if len(stripped) > 80 else "")
-            warnings.append(f"Malformed bullet: {short}")
-    return warnings
-
-
-def _validate_diff(old_memory: str | None, new_memory: str) -> list[str]:
-    """Return a list of warning strings about suspicious changes."""
-    if not old_memory:
-        return []
-
-    warnings: list[str] = []
-    old_headings = _extract_headings(old_memory)
-    new_headings = _extract_headings(new_memory)
-    dropped = old_headings - new_headings
-    if dropped:
-        names = ", ".join(sorted(dropped))
-        warnings.append(
-            f"Sections removed: {names}. "
-            "If unintentional, the user can restore from the settings page."
-        )
-
-    old_len = len(old_memory)
-    new_len = len(new_memory)
-    if old_len > 0 and new_len < old_len * 0.4:
-        warnings.append(
-            f"Memory shrank significantly ({old_len:,} -> {new_len:,} chars). "
-            "Possible data loss."
-        )
-    return warnings
-
-
-# ---------------------------------------------------------------------------
-# Size validation & soft warning
-# ---------------------------------------------------------------------------
-
-
-def _validate_memory_size(content: str) -> dict[str, Any] | None:
-    """Return an error/warning dict if *content* is too large, else None."""
-    length = len(content)
-    if length > MEMORY_HARD_LIMIT:
-        return {
-            "status": "error",
-            "message": (
-                f"Memory exceeds {MEMORY_HARD_LIMIT:,} character limit "
-                f"({length:,} chars). Consolidate by merging related items, "
-                "removing outdated entries, and shortening descriptions. "
-                "Then call update_memory again."
-            ),
-        }
-    return None
-
-
-def _soft_warning(content: str) -> str | None:
-    """Return a warning string if content exceeds the soft limit."""
-    length = len(content)
-    if length > MEMORY_SOFT_LIMIT:
-        return (
-            f"Memory is at {length:,}/{MEMORY_HARD_LIMIT:,} characters. "
-            "Consolidate by merging related items and removing less important "
-            "entries on your next update."
-        )
-    return None
-
-
-# ---------------------------------------------------------------------------
-# Forced rewrite when memory exceeds the hard limit
-# ---------------------------------------------------------------------------
-
-_FORCED_REWRITE_PROMPT = """\
-You are a memory curator. The following memory document exceeds the character \
-limit and must be shortened.
-
-RULES:
-1. Rewrite the document to be under {target} characters.
-2. Preserve existing ## headings. Every entry must remain under a heading. You may merge
-   or rename headings to consolidate, but keep names personal and descriptive.
-3. Priority for keeping content: [instr] > [pref] > [fact].
-4. Merge duplicate entries, remove outdated entries, shorten verbose descriptions.
-5. Every bullet MUST have format: - (YYYY-MM-DD) [fact|pref|instr] text
-6. Preserve the user's first name in entries — do not replace it with "the user".
-7. Output ONLY the consolidated markdown — no explanations, no wrapping.
-
-<memory_document>
-{content}
-</memory_document>"""
-
-
-async def _forced_rewrite(content: str, llm: Any) -> str | None:
-    """Use a focused LLM call to compress *content* under the hard limit.
-
-    Returns the rewritten string, or ``None`` if the call fails.
-    """
-    try:
-        prompt = _FORCED_REWRITE_PROMPT.format(
-            target=MEMORY_HARD_LIMIT, content=content
-        )
-        response = await llm.ainvoke(
-            [HumanMessage(content=prompt)],
-            config={"tags": ["surfsense:internal"]},
-        )
-        text = extract_text_content(response.content).strip()
-        if not text:
-            logger.warning("Forced rewrite returned empty text; aborting rewrite")
-            return None
-        return text
-    except Exception:
-        logger.exception("Forced rewrite LLM call failed")
-        return None
-
-
-# ---------------------------------------------------------------------------
-# Shared save-and-respond logic
-# ---------------------------------------------------------------------------
-
-
-async def _save_memory(
-    *,
-    updated_memory: str,
-    old_memory: str | None,
-    llm: Any | None,
-    apply_fn,
-    commit_fn,
-    rollback_fn,
-    label: str,
-    scope: Literal["user", "team"],
-) -> dict[str, Any]:
-    """Validate, optionally force-rewrite if over the hard limit, save, and
-    return a response dict.
-
-    Parameters
-    ----------
-    updated_memory : str
-        The new document the agent submitted.
-    old_memory : str | None
-        The previously persisted document (for diff checks).
-    llm : Any | None
-        LLM instance for forced rewrite (may be ``None``).
-    apply_fn : callable(str) -> None
-        Callback that sets the new memory on the ORM object.
-    commit_fn : coroutine
-        ``session.commit``.
-    rollback_fn : coroutine
-        ``session.rollback``.
-    label : str
-        Human label for log messages (e.g. "user memory", "team memory").
-    """
-    if not isinstance(updated_memory, str):
-        logger.warning(
-            "Refusing non-string memory payload (type=%s)",
-            type(updated_memory).__name__,
-        )
-        return {
-            "status": "error",
-            "message": "Internal error: memory payload must be a string.",
-        }
-
-    content = updated_memory
-
-    # --- forced rewrite if over the hard limit ---
-    if len(content) > MEMORY_HARD_LIMIT and llm is not None:
-        rewritten = await _forced_rewrite(content, llm)
-        if rewritten is not None and len(rewritten) < len(content):
-            content = rewritten
-
-    # --- hard-limit gate (reject if still too large after rewrite) ---
-    size_err = _validate_memory_size(content)
-    if size_err:
-        return size_err
-
-    scope_err = _validate_memory_scope(content, scope)
-    if scope_err:
-        return scope_err
-
-    # --- persist ---
-    try:
-        apply_fn(content)
-        await commit_fn()
-    except Exception as e:
-        logger.exception("Failed to update %s: %s", label, e)
-        await rollback_fn()
-        return {"status": "error", "message": f"Failed to update {label}: {e}"}
-
-    # --- build response ---
-    resp: dict[str, Any] = {
-        "status": "saved",
-        "message": f"{label.capitalize()} updated.",
-    }
-
-    if content is not updated_memory:
-        resp["notice"] = "Memory was automatically rewritten to fit within limits."
-
-    diff_warnings = _validate_diff(old_memory, content)
-    if diff_warnings:
-        resp["diff_warnings"] = diff_warnings
-
-    format_warnings = _validate_bullet_format(content)
-    if format_warnings:
-        resp["format_warnings"] = format_warnings
-
-    warning = _soft_warning(content)
-    if warning:
-        resp["warning"] = warning
-
-    return resp
-
-
-# ---------------------------------------------------------------------------
-# Tool factories
-# ---------------------------------------------------------------------------
-
 
 def create_update_memory_tool(
     user_id: str | UUID,
     db_session: AsyncSession,
     llm: Any | None = None,
 ):
-    """Factory function to create the user-memory update tool.
+    """Factory for the user-memory update tool.
 
-    The tool acquires its own short-lived ``AsyncSession`` per call via
-    :data:`async_session_maker` so the closure is safe to share across
-    HTTP requests by the compiled-agent cache. Capturing a per-request
-    session here would surface stale/closed sessions on cache hits.
-    The session's bound ``commit``/``rollback`` methods are captured at
-    call time, after ``async with`` has bound ``db_session`` locally.
-
-    Args:
-        user_id: ID of the user whose memory document is being updated.
-        db_session: Reserved for registry compatibility. Per-call sessions
-            are opened via :data:`async_session_maker` inside the tool body.
-        llm: Optional LLM for the forced-rewrite path.
-
-    Returns:
-        Configured update_memory tool for the user-memory scope.
+    Uses a fresh short-lived session per call so compiled-agent caches never
+    retain a stale request-scoped session.
     """
-    del db_session  # per-call session — see docstring
+    del db_session
     uid = UUID(user_id) if isinstance(user_id, str) else user_id
 
     @tool
     async def update_memory(updated_memory: str) -> dict[str, Any]:
         """Update the user's personal memory document.
 
-        Your current memory is shown in <user_memory> in the system prompt.
-        When the user shares important long-term information (preferences,
-        facts, instructions, context), rewrite the memory document to include
-        the new information.  Merge new facts with existing ones, update
-        contradictions, remove outdated entries, and keep it concise.
-
-        Args:
-            updated_memory: The FULL updated markdown document (not a diff).
+        The current memory is shown in <user_memory>. Pass the FULL updated
+        markdown document, not a diff.
         """
         try:
             async with async_session_maker() as db_session:
-                result = await db_session.execute(select(User).where(User.id == uid))
-                user = result.scalars().first()
-                if not user:
-                    return {"status": "error", "message": "User not found."}
-
-                old_memory = user.memory_md
-
-                return await _save_memory(
-                    updated_memory=updated_memory,
-                    old_memory=old_memory,
+                result = await save_memory(
+                    scope=MemoryScope.USER,
+                    target_id=uid,
+                    content=updated_memory,
+                    session=db_session,
                     llm=llm,
-                    apply_fn=lambda content: setattr(user, "memory_md", content),
-                    commit_fn=db_session.commit,
-                    rollback_fn=db_session.rollback,
-                    label="memory",
-                    scope="user",
                 )
+                return result.to_dict()
         except Exception as e:
             logger.exception("Failed to update user memory: %s", e)
-            return {
-                "status": "error",
-                "message": f"Failed to update memory: {e}",
-            }
+            return {"status": "error", "message": f"Failed to update memory: {e}"}
 
     return update_memory
 
@@ -373,64 +57,26 @@ def create_update_team_memory_tool(
     db_session: AsyncSession,
     llm: Any | None = None,
 ):
-    """Factory function to create the team-memory update tool.
-
-    The tool acquires its own short-lived ``AsyncSession`` per call via
-    :data:`async_session_maker` so the closure is safe to share across
-    HTTP requests by the compiled-agent cache. Capturing a per-request
-    session here would surface stale/closed sessions on cache hits.
-    The session's bound ``commit``/``rollback`` methods are captured at
-    call time, after ``async with`` has bound ``db_session`` locally.
-
-    Args:
-        search_space_id: ID of the search space whose team memory is being
-            updated.
-        db_session: Reserved for registry compatibility. Per-call sessions
-            are opened via :data:`async_session_maker` inside the tool body.
-        llm: Optional LLM for the forced-rewrite path.
-
-    Returns:
-        Configured update_memory tool for the team-memory scope.
-    """
-    del db_session  # per-call session — see docstring
+    """Factory for the team-memory update tool."""
+    del db_session
 
     @tool
     async def update_memory(updated_memory: str) -> dict[str, Any]:
         """Update the team's shared memory document for this search space.
 
-        Your current team memory is shown in <team_memory> in the system
-        prompt.  When the team shares important long-term information
-        (decisions, conventions, key facts, priorities), rewrite the memory
-        document to include the new information.  Merge new facts with
-        existing ones, update contradictions, remove outdated entries, and
-        keep it concise.
-
-        Args:
-            updated_memory: The FULL updated markdown document (not a diff).
+        The current team memory is shown in <team_memory>. Pass the FULL updated
+        markdown document, not a diff.
         """
         try:
             async with async_session_maker() as db_session:
-                result = await db_session.execute(
-                    select(SearchSpace).where(SearchSpace.id == search_space_id)
-                )
-                space = result.scalars().first()
-                if not space:
-                    return {"status": "error", "message": "Search space not found."}
-
-                old_memory = space.shared_memory_md
-
-                return await _save_memory(
-                    updated_memory=updated_memory,
-                    old_memory=old_memory,
+                result = await save_memory(
+                    scope=MemoryScope.TEAM,
+                    target_id=search_space_id,
+                    content=updated_memory,
+                    session=db_session,
                     llm=llm,
-                    apply_fn=lambda content: setattr(
-                        space, "shared_memory_md", content
-                    ),
-                    commit_fn=db_session.commit,
-                    rollback_fn=db_session.rollback,
-                    label="team memory",
-                    scope="team",
                 )
+                return result.to_dict()
         except Exception as e:
             logger.exception("Failed to update team memory: %s", e)
             return {
@@ -439,3 +85,9 @@ def create_update_team_memory_tool(
             }
 
     return update_memory
+
+
+__all__ = [
+    "create_update_memory_tool",
+    "create_update_team_memory_tool",
+]

surfsense_backend/app/config/__init__.py

@@ -110,6 +110,19 @@ def load_global_llm_configs():
         except Exception as e:
             print(f"Warning: Failed to score global LLM configs: {e}")
 
+        # Planner LLM is a singleton role. If an operator accidentally
+        # marks multiple configs ``is_planner: true``, only the first one
+        # is used at runtime — surface the others at startup so the
+        # mistake is caught before traffic, not silently buried.
+        planner_cfgs = [c for c in configs if c.get("is_planner") is True]
+        if len(planner_cfgs) > 1:
+            extra_ids = [c.get("id") for c in planner_cfgs[1:]]
+            print(
+                "Warning: Multiple global LLM configs marked is_planner=true "
+                f"(ids {[c.get('id') for c in planner_cfgs]}); using id "
+                f"{planner_cfgs[0].get('id')} and ignoring {extra_ids}"
+            )
+
         return configs
     except Exception as e:
         print(f"Warning: Failed to load global LLM configs: {e}")

surfsense_backend/app/config/global_llm_config.example.yaml

@@ -258,6 +258,45 @@ global_llm_configs:
     use_default_system_instructions: true
     citations_enabled: true
 
+  # Example: Planner LLM - small, fast model used for internal utility tasks
+  #
+  # The PLANNER role handles short, structured internal calls (KB query
+  # rewriting, date extraction, recency classification, etc.) that don't
+  # need frontier-tier capability. Pointing the planner at a cheap+fast
+  # model (gpt-4o-mini, Claude Haiku, Azure gpt-5.x-nano, Groq Llama, ...)
+  # typically saves 500ms-1.5s per turn vs. routing those same internal
+  # calls through the user's chat model.
+  #
+  # Activation:
+  #   - Mark EXACTLY ONE global config with ``is_planner: true``.
+  #   - If multiple are marked, the first one wins and a WARNING is logged.
+  #   - If none is marked, every internal call falls back to the user's
+  #     chat LLM (same behavior as before this flag existed).
+  #
+  # This config is operator-only — it is NOT exposed in the user-facing
+  # model selector, never billed against premium quota, and the
+  # billing_tier / anonymous_enabled fields below are ignored.
+  - id: -9
+    name: "Global Planner (GPT-4o mini)"
+    description: "Internal-only planner LLM for query rewriting and classification"
+    is_planner: true
+    billing_tier: "free"
+    anonymous_enabled: false
+    seo_enabled: false
+    quota_reserve_tokens: 1000
+    provider: "OPENAI"
+    model_name: "gpt-4o-mini"
+    api_key: "sk-your-openai-api-key-here"
+    api_base: ""
+    rpm: 3500
+    tpm: 200000
+    litellm_params:
+      temperature: 0
+      max_tokens: 1000
+    system_instructions: ""
+    use_default_system_instructions: true
+    citations_enabled: false
+
 # =============================================================================
 # OpenRouter Integration
 # =============================================================================
@@ -493,6 +532,20 @@ global_vision_llm_configs:
 # - Lower temperature (0.3) is recommended for accurate screenshot analysis
 # - Lower max_tokens (1000) is sufficient since autocomplete produces short suggestions
 #
+# PLANNER LLM NOTES:
+# - is_planner: true marks a config as the internal-only planner LLM (small,
+#   fast model used for KB query rewriting, date extraction, recency
+#   classification, etc.). Only one config may carry this flag — if
+#   multiple do, the first one wins and a startup WARNING is logged.
+# - When no config is marked is_planner, every internal utility call falls
+#   back to the user's chat LLM (the historical behavior).
+# - Planner configs are NOT shown in the user-facing model selector and
+#   are NOT billed against the user's premium quota. Their billing_tier,
+#   anonymous_enabled, seo_* fields are ignored.
+# - Recommended models: gpt-4o-mini, claude-3-5-haiku, gemini-1.5-flash,
+#   azure gpt-5.x-nano, groq llama3-8b — anything <200ms p50 on a 1-2k
+#   prompt. Frontier models here defeat the purpose of the flag.
+#
 # TOKEN QUOTA & ANONYMOUS ACCESS NOTES:
 # - billing_tier: "free" or "premium". Controls whether registered users need premium token quota.
 # - anonymous_enabled: true/false. Whether the model appears in the public no-login catalog.

surfsense_backend/app/routes/__init__.py

@@ -54,6 +54,7 @@ from .search_spaces_routes import router as search_spaces_router
 from .slack_add_connector_route import router as slack_add_connector_router
 from .stripe_routes import router as stripe_router
 from .surfsense_docs_routes import router as surfsense_docs_router
+from .team_memory_routes import router as team_memory_router
 from .teams_add_connector_route import router as teams_add_connector_router
 from .video_presentations_routes import router as video_presentations_router
 from .vision_llm_routes import router as vision_llm_router
@@ -117,3 +118,4 @@ router.include_router(stripe_router)  # Stripe checkout for additional page pack
 router.include_router(youtube_router)  # YouTube playlist resolution
 router.include_router(prompts_router)
 router.include_router(memory_router)  # User personal memory (memory.md style)
+router.include_router(team_memory_router)  # Search-space team memory

surfsense_backend/app/routes/mcp_oauth_route.py

@@ -428,7 +428,7 @@ async def mcp_oauth_callback(
             await session.commit()
             await session.refresh(db_connector)
 
-            _invalidate_cache(space_id)
+            _refresh_mcp_cache(db_connector.id, space_id)
 
             logger.info(
                 "Re-authenticated %s MCP connector %s for user %s",
@@ -481,7 +481,7 @@ async def mcp_oauth_callback(
                 detail="A connector for this service already exists.",
             ) from e
 
-        _invalidate_cache(space_id)
+        _refresh_mcp_cache(new_connector.id, space_id)
 
         logger.info(
             "Created %s MCP connector %s for user %s in space %s",
@@ -658,10 +658,17 @@ async def reauth_mcp_service(
 # ---------------------------------------------------------------------------
 
 
-def _invalidate_cache(space_id: int) -> None:
-    try:
-        from app.agents.new_chat.tools.mcp_tool import invalidate_mcp_tools_cache
+def _refresh_mcp_cache(connector_id: int, space_id: int) -> None:
+    """Evict the in-process MCP tool LRU and schedule background prefetch.
 
-        invalidate_mcp_tools_cache(space_id)
+    Wraps :func:`refresh_mcp_tools_cache_for_connector` so any failure is
+    isolated from the OAuth response flow.
+    """
+    try:
+        from app.agents.new_chat.tools.mcp_tools_cache import (
+            refresh_mcp_tools_cache_for_connector,
+        )
+
+        refresh_mcp_tools_cache_for_connector(connector_id, space_id)
     except Exception:
-        logger.debug("MCP cache invalidation skipped", exc_info=True)
+        logger.debug("MCP cache refresh skipped", exc_info=True)

surfsense_backend/app/routes/memory_routes.py

@@ -1,75 +1,40 @@
-"""Routes for user memory management (personal memory.md)."""
+"""Routes for user memory management."""
 
 from __future__ import annotations
 
-import logging
-
 from fastapi import APIRouter, Depends, HTTPException
-from langchain_core.messages import HumanMessage
 from pydantic import BaseModel
 from sqlalchemy.ext.asyncio import AsyncSession
 
-from app.agents.new_chat.llm_config import (
-    create_chat_litellm_from_agent_config,
-    load_agent_llm_config_for_search_space,
-)
-from app.agents.new_chat.tools.update_memory import MEMORY_HARD_LIMIT, _save_memory
 from app.db import User, get_async_session
+from app.services.memory import (
+    MemoryRead,
+    MemoryScope,
+    memory_limits,
+    read_memory,
+    reset_memory,
+    save_memory,
+)
 from app.users import current_active_user
-from app.utils.content_utils import extract_text_content
-
-logger = logging.getLogger(__name__)
 
 router = APIRouter()
 
 
-class MemoryRead(BaseModel):
-    memory_md: str
-
-
 class MemoryUpdate(BaseModel):
     memory_md: str
 
 
-class MemoryEditRequest(BaseModel):
-    query: str
-    search_space_id: int
-
-
-_MEMORY_EDIT_PROMPT = """\
-You are a memory editor. The user wants to modify their memory document. \
-Apply the user's instruction to the existing memory document and output the \
-FULL updated document.
-
-RULES:
-1. If the instruction asks to add something, add it with format: \
-- (YYYY-MM-DD) [fact|pref|instr] text, under an existing or new ## heading. \
-Heading names should be personal and descriptive, not generic categories.
-2. If the instruction asks to remove something, remove the matching entry.
-3. If the instruction asks to change something, update the matching entry.
-4. Preserve existing ## headings and all other entries.
-5. Every bullet must include a marker: [fact], [pref], or [instr].
-6. Use the user's first name (from <user_name>) in entries instead of "the user".
-7. Output ONLY the updated markdown — no explanations, no wrapping.
-
-<user_name>{user_name}</user_name>
-
-<current_memory>
-{current_memory}
-</current_memory>
-
-<user_instruction>
-{instruction}
-</user_instruction>"""
-
-
 @router.get("/users/me/memory", response_model=MemoryRead)
 async def get_user_memory(
     user: User = Depends(current_active_user),
     session: AsyncSession = Depends(get_async_session),
 ):
-    await session.refresh(user, ["memory_md"])
-    return MemoryRead(memory_md=user.memory_md or "")
+    memory_md = await read_memory(
+        scope=MemoryScope.USER,
+        target_id=user.id,
+        session=session,
+    )
+    return MemoryRead(memory_md=memory_md, limits=memory_limits())
 
 
 @router.put("/users/me/memory", response_model=MemoryRead)
@@ -78,73 +43,27 @@ async def update_user_memory(
     user: User = Depends(current_active_user),
     session: AsyncSession = Depends(get_async_session),
 ):
-    if len(body.memory_md) > MEMORY_HARD_LIMIT:
-        raise HTTPException(
-            status_code=400,
-            detail=f"Memory exceeds {MEMORY_HARD_LIMIT:,} character limit ({len(body.memory_md):,} chars).",
-        )
-    user.memory_md = body.memory_md
-    session.add(user)
-    await session.commit()
-    await session.refresh(user, ["memory_md"])
-    return MemoryRead(memory_md=user.memory_md or "")
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id=user.id,
+        content=body.memory_md,
+        session=session,
+    )
+    if result.status == "error":
+        raise HTTPException(status_code=400, detail=result.message)
+    return MemoryRead(memory_md=result.memory_md, limits=memory_limits())
 
 
-@router.post("/users/me/memory/edit", response_model=MemoryRead)
-async def edit_user_memory(
-    body: MemoryEditRequest,
+@router.post("/users/me/memory/reset", response_model=MemoryRead)
+async def reset_user_memory(
     user: User = Depends(current_active_user),
     session: AsyncSession = Depends(get_async_session),
 ):
-    """Apply a natural language edit to the user's personal memory via LLM."""
-    agent_config = await load_agent_llm_config_for_search_space(
-        session, body.search_space_id
+    result = await reset_memory(
+        scope=MemoryScope.USER,
+        target_id=user.id,
+        session=session,
     )
-    if not agent_config:
-        raise HTTPException(status_code=500, detail="No LLM configuration available.")
-    llm = create_chat_litellm_from_agent_config(agent_config)
-    if not llm:
-        raise HTTPException(status_code=500, detail="Failed to create LLM instance.")
-
-    await session.refresh(user, ["memory_md", "display_name"])
-    current_memory = user.memory_md or ""
-    first_name = (
-        user.display_name.strip().split()[0]
-        if user.display_name and user.display_name.strip()
-        else "The user"
-    )
-
-    prompt = _MEMORY_EDIT_PROMPT.format(
-        current_memory=current_memory or "(empty)",
-        instruction=body.query,
-        user_name=first_name,
-    )
-    try:
-        response = await llm.ainvoke(
-            [HumanMessage(content=prompt)],
-            config={"tags": ["surfsense:internal", "memory-edit"]},
-        )
-        updated = extract_text_content(response.content).strip()
-    except Exception as e:
-        logger.exception("Memory edit LLM call failed: %s", e)
-        raise HTTPException(status_code=500, detail="Memory edit failed.") from e
-
-    if not updated:
-        raise HTTPException(status_code=400, detail="LLM returned empty result.")
-
-    result = await _save_memory(
-        updated_memory=updated,
-        old_memory=current_memory,
-        llm=llm,
-        apply_fn=lambda content: setattr(user, "memory_md", content),
-        commit_fn=session.commit,
-        rollback_fn=session.rollback,
-        label="memory",
-        scope="user",
-    )
-
-    if result.get("status") == "error":
-        raise HTTPException(status_code=400, detail=result["message"])
-
-    await session.refresh(user, ["memory_md"])
-    return MemoryRead(memory_md=user.memory_md or "")
+    if result.status == "error":
+        raise HTTPException(status_code=400, detail=result.message)
+    return MemoryRead(memory_md=result.memory_md, limits=memory_limits())

surfsense_backend/app/routes/search_source_connectors_routes.py

@@ -2650,9 +2650,11 @@ async def create_mcp_connector(
             f"for user {user.id} in search space {search_space_id}"
         )
 
-        from app.agents.new_chat.tools.mcp_tool import invalidate_mcp_tools_cache
+        from app.agents.new_chat.tools.mcp_tools_cache import (
+            refresh_mcp_tools_cache_for_connector,
+        )
 
-        invalidate_mcp_tools_cache(search_space_id)
+        refresh_mcp_tools_cache_for_connector(db_connector.id, search_space_id)
 
         connector_read = SearchSourceConnectorRead.model_validate(db_connector)
         return MCPConnectorRead.from_connector(connector_read)
@@ -2828,9 +2830,11 @@ async def update_mcp_connector(
 
         logger.info(f"Updated MCP connector {connector_id}")
 
-        from app.agents.new_chat.tools.mcp_tool import invalidate_mcp_tools_cache
+        from app.agents.new_chat.tools.mcp_tools_cache import (
+            refresh_mcp_tools_cache_for_connector,
+        )
 
-        invalidate_mcp_tools_cache(connector.search_space_id)
+        refresh_mcp_tools_cache_for_connector(connector.id, connector.search_space_id)
 
         connector_read = SearchSourceConnectorRead.model_validate(connector)
         return MCPConnectorRead.from_connector(connector_read)

surfsense_backend/app/routes/search_spaces_routes.py

@@ -1,17 +1,10 @@
 import logging
 
 from fastapi import APIRouter, Depends, HTTPException
-from langchain_core.messages import HumanMessage
-from pydantic import BaseModel as PydanticBaseModel
 from sqlalchemy import func, update
 from sqlalchemy.ext.asyncio import AsyncSession
 from sqlalchemy.future import select
 
-from app.agents.new_chat.llm_config import (
-    create_chat_litellm_from_agent_config,
-    load_agent_llm_config_for_search_space,
-)
-from app.agents.new_chat.tools.update_memory import MEMORY_HARD_LIMIT, _save_memory
 from app.config import config
 from app.db import (
     ImageGenerationConfig,
@@ -35,7 +28,6 @@ from app.schemas import (
     SearchSpaceWithStats,
 )
 from app.users import current_active_user
-from app.utils.content_utils import extract_text_content
 from app.utils.rbac import check_permission, check_search_space_access
 
 logger = logging.getLogger(__name__)
@@ -43,34 +35,6 @@ logger = logging.getLogger(__name__)
 router = APIRouter()
 
 
-class _TeamMemoryEditRequest(PydanticBaseModel):
-    query: str
-
-
-_TEAM_MEMORY_EDIT_PROMPT = """\
-You are a memory editor for a team workspace. The user wants to modify the \
-team's shared memory document. Apply the user's instruction to the existing \
-memory document and output the FULL updated document.
-
-RULES:
-1. If the instruction asks to add something, add it with format: \
-- (YYYY-MM-DD) [fact] text, under an existing or new ## heading. \
-Heading names should be descriptive, not generic categories.
-2. If the instruction asks to remove something, remove the matching entry.
-3. If the instruction asks to change something, update the matching entry.
-4. Preserve existing ## headings and all other entries.
-5. NEVER use [pref] or [instr] markers. Team memory uses [fact] only.
-6. Output ONLY the updated markdown — no explanations, no wrapping.
-
-<current_memory>
-{current_memory}
-</current_memory>
-
-<user_instruction>
-{instruction}
-</user_instruction>"""
-
-
 async def create_default_roles_and_membership(
     session: AsyncSession,
     search_space_id: int,
@@ -294,15 +258,6 @@ async def update_search_space(
 
         update_data = search_space_update.model_dump(exclude_unset=True)
 
-        if (
-            "shared_memory_md" in update_data
-            and len(update_data["shared_memory_md"] or "") > MEMORY_HARD_LIMIT
-        ):
-            raise HTTPException(
-                status_code=400,
-                detail=f"Team memory exceeds {MEMORY_HARD_LIMIT:,} character limit.",
-            )
-
         for key, value in update_data.items():
             setattr(db_search_space, key, value)
         await session.commit()
@@ -317,72 +272,6 @@ async def update_search_space(
         ) from e
 
 
-@router.post(
-    "/searchspaces/{search_space_id}/memory/edit",
-    response_model=SearchSpaceRead,
-)
-async def edit_team_memory(
-    search_space_id: int,
-    body: _TeamMemoryEditRequest,
-    session: AsyncSession = Depends(get_async_session),
-    user: User = Depends(current_active_user),
-):
-    """Apply a natural language edit to the team memory via LLM."""
-    await check_search_space_access(session, user, search_space_id)
-
-    agent_config = await load_agent_llm_config_for_search_space(
-        session, search_space_id
-    )
-    if not agent_config:
-        raise HTTPException(status_code=500, detail="No LLM configuration available.")
-    llm = create_chat_litellm_from_agent_config(agent_config)
-    if not llm:
-        raise HTTPException(status_code=500, detail="Failed to create LLM instance.")
-
-    result = await session.execute(
-        select(SearchSpace).filter(SearchSpace.id == search_space_id)
-    )
-    db_search_space = result.scalars().first()
-    if not db_search_space:
-        raise HTTPException(status_code=404, detail="Search space not found")
-
-    current_memory = db_search_space.shared_memory_md or ""
-
-    prompt = _TEAM_MEMORY_EDIT_PROMPT.format(
-        current_memory=current_memory or "(empty)",
-        instruction=body.query,
-    )
-    try:
-        response = await llm.ainvoke(
-            [HumanMessage(content=prompt)],
-            config={"tags": ["surfsense:internal", "memory-edit"]},
-        )
-        updated = extract_text_content(response.content).strip()
-    except Exception as e:
-        logger.exception("Team memory edit LLM call failed: %s", e)
-        raise HTTPException(status_code=500, detail="Team memory edit failed.") from e
-
-    if not updated:
-        raise HTTPException(status_code=400, detail="LLM returned empty result.")
-
-    save_result = await _save_memory(
-        updated_memory=updated,
-        old_memory=current_memory,
-        llm=llm,
-        apply_fn=lambda content: setattr(db_search_space, "shared_memory_md", content),
-        commit_fn=session.commit,
-        rollback_fn=session.rollback,
-        label="team memory",
-        scope="team",
-    )
-
-    if save_result.get("status") == "error":
-        raise HTTPException(status_code=400, detail=save_result["message"])
-
-    await session.refresh(db_search_space)
-    return db_search_space
-
-
 @router.post("/searchspaces/{search_space_id}/ai-sort")
 async def trigger_ai_sort(
     search_space_id: int,

surfsense_backend/app/routes/team_memory_routes.py

@@ -0,0 +1,76 @@
+"""Routes for search-space team memory."""
+
+from __future__ import annotations
+
+from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.db import User, get_async_session
+from app.services.memory import (
+    MemoryRead,
+    MemoryScope,
+    memory_limits,
+    read_memory,
+    reset_memory,
+    save_memory,
+)
+from app.users import current_active_user
+from app.utils.rbac import check_search_space_access
+
+router = APIRouter()
+
+
+class TeamMemoryUpdate(BaseModel):
+    memory_md: str
+
+
+@router.get("/searchspaces/{search_space_id}/memory", response_model=MemoryRead)
+async def get_team_memory(
+    search_space_id: int,
+    session: AsyncSession = Depends(get_async_session),
+    user: User = Depends(current_active_user),
+):
+    await check_search_space_access(session, user, search_space_id)
+    memory_md = await read_memory(
+        scope=MemoryScope.TEAM,
+        target_id=search_space_id,
+        session=session,
+    )
+    return MemoryRead(memory_md=memory_md, limits=memory_limits())
+
+
+@router.put("/searchspaces/{search_space_id}/memory", response_model=MemoryRead)
+async def update_team_memory(
+    search_space_id: int,
+    body: TeamMemoryUpdate,
+    session: AsyncSession = Depends(get_async_session),
+    user: User = Depends(current_active_user),
+):
+    await check_search_space_access(session, user, search_space_id)
+    result = await save_memory(
+        scope=MemoryScope.TEAM,
+        target_id=search_space_id,
+        content=body.memory_md,
+        session=session,
+    )
+    if result.status == "error":
+        raise HTTPException(status_code=400, detail=result.message)
+    return MemoryRead(memory_md=result.memory_md, limits=memory_limits())
+
+
+@router.post("/searchspaces/{search_space_id}/memory/reset", response_model=MemoryRead)
+async def reset_team_memory(
+    search_space_id: int,
+    session: AsyncSession = Depends(get_async_session),
+    user: User = Depends(current_active_user),
+):
+    await check_search_space_access(session, user, search_space_id)
+    result = await reset_memory(
+        scope=MemoryScope.TEAM,
+        target_id=search_space_id,
+        session=session,
+    )
+    if result.status == "error":
+        raise HTTPException(status_code=400, detail=result.message)
+    return MemoryRead(memory_md=result.memory_md, limits=memory_limits())

surfsense_backend/app/schemas/search_space.py

@@ -21,7 +21,6 @@ class SearchSpaceUpdate(BaseModel):
     description: str | None = None
     citations_enabled: bool | None = None
     qna_custom_instructions: str | None = None
-    shared_memory_md: str | None = None
     ai_file_sort_enabled: bool | None = None
 
 

surfsense_backend/app/services/gmail/kb_sync_service.py

@@ -1,3 +1,4 @@
+import asyncio
 import logging
 from datetime import datetime
 
@@ -100,7 +101,9 @@ class GmailKBSyncService:
             else:
                 logger.warning("No LLM configured -- using fallback summary")
                 summary_content = f"Gmail Message: {subject}\n\n{indexable_content}"
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(indexable_content)
             now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

surfsense_backend/app/services/google_calendar/kb_sync_service.py

@@ -116,7 +116,9 @@ class GoogleCalendarKBSyncService:
                 summary_content = (
                     f"Google Calendar Event: {event_summary}\n\n{indexable_content}"
                 )
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(indexable_content)
             now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
@@ -295,7 +297,9 @@ class GoogleCalendarKBSyncService:
                 summary_content = (
                     f"Google Calendar Event: {event_summary}\n\n{indexable_content}"
                 )
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(indexable_content)
             now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

surfsense_backend/app/services/jira/kb_sync_service.py

@@ -98,7 +98,9 @@ class JiraKBSyncService:
                 summary_content = (
                     f"Jira Issue {issue_identifier}: {issue_title}\n\n{issue_content}"
                 )
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(issue_content)
             now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
@@ -212,7 +214,9 @@ class JiraKBSyncService:
                 summary_content = (
                     f"Jira Issue {issue_identifier}: {issue_title}\n\n{issue_content}"
                 )
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(issue_content)
 

surfsense_backend/app/services/llm_service.py

@@ -659,3 +659,36 @@ async def get_user_long_context_llm(
     return await get_document_summary_llm(
         session, search_space_id, disable_streaming=disable_streaming
     )
+
+
+def get_planner_llm() -> ChatLiteLLM | None:
+    """Return a planner LLM instance from the first global config marked
+    ``is_planner: true``, or ``None`` if no planner config is defined.
+
+    The planner role handles short, structured internal tasks (KB search
+    planning: query rewriting, date extraction, recency classification).
+    These tasks are well-served by small/fast models (e.g. gpt-4o-mini,
+    Claude Haiku, Azure gpt-5.x-nano) — using the user's chat LLM for them
+    is unnecessarily expensive and slow.
+
+    This helper reads from ``config.GLOBAL_LLM_CONFIGS`` (loaded at import
+    time from ``global_llm_config.yaml``) so it has no DB cost and can be
+    called synchronously from middleware/factory code. It returns the same
+    instance shape as the global path of ``get_search_space_llm_instance``.
+
+    Callers MUST fall back to their chat LLM when this returns ``None`` so
+    deployments without a planner config keep working unchanged.
+    """
+    from app.agents.new_chat.llm_config import create_chat_litellm_from_config
+
+    planner_cfg = next(
+        (
+            cfg
+            for cfg in config.GLOBAL_LLM_CONFIGS
+            if cfg.get("is_planner") is True
+        ),
+        None,
+    )
+    if not planner_cfg:
+        return None
+    return create_chat_litellm_from_config(planner_cfg)

surfsense_backend/app/services/memory/__init__.py

@@ -0,0 +1,32 @@
+"""First-class memory service for user and team markdown memory."""
+
+from .schemas import MemoryLimits, MemoryRead
+from .service import (
+    MemoryScope,
+    SaveResult,
+    memory_limits,
+    read_memory,
+    reset_memory,
+    save_memory,
+)
+from .validation import (
+    MEMORY_HARD_LIMIT,
+    MEMORY_SOFT_LIMIT,
+    validate_bullet_format,
+    validate_memory_scope,
+)
+
+__all__ = [
+    "MEMORY_HARD_LIMIT",
+    "MEMORY_SOFT_LIMIT",
+    "MemoryLimits",
+    "MemoryRead",
+    "MemoryScope",
+    "SaveResult",
+    "memory_limits",
+    "read_memory",
+    "reset_memory",
+    "save_memory",
+    "validate_bullet_format",
+    "validate_memory_scope",
+]

surfsense_backend/app/services/memory/document.py

@@ -0,0 +1,200 @@
+"""Memory-specific markdown document model and canonical renderer.
+
+This intentionally parses only SurfSense memory's small markdown contract:
+``##`` sections with dated bullet items. Unknown lines are preserved so user
+edits are not lost, while legacy marker bullets are normalized on render.
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from datetime import date
+
+DEFAULT_LEGACY_SECTION = "Memory"
+LEGACY_MARKERS = frozenset({"fact", "pref", "instr"})
+
+
+@dataclass(frozen=True)
+class MemoryBullet:
+    entry_date: date
+    text: str
+
+
+@dataclass(frozen=True)
+class MemoryRawLine:
+    text: str
+
+
+MemoryLine = MemoryBullet | MemoryRawLine
+
+
+@dataclass(frozen=True)
+class MemorySection:
+    heading: str
+    lines: list[MemoryLine] = field(default_factory=list)
+    explicit_heading: bool = True
+
+
+@dataclass(frozen=True)
+class MemoryDocument:
+    sections: list[MemorySection] = field(default_factory=list)
+
+    @property
+    def has_explicit_heading(self) -> bool:
+        return any(section.explicit_heading for section in self.sections)
+
+
+def is_section_heading(line: str) -> bool:
+    return line.startswith("## ") and bool(line[3:].strip())
+
+
+def heading_text(line: str) -> str:
+    return line[3:].strip()
+
+
+def normalize_heading(heading: str) -> str:
+    chars: list[str] = []
+    previous_was_space = True
+    for char in heading.strip().lower():
+        if char.isalnum():
+            chars.append(char)
+            previous_was_space = False
+        elif not previous_was_space:
+            chars.append(" ")
+            previous_was_space = True
+    return "".join(chars).strip()
+
+
+def parse_bullet_line(line: str) -> MemoryBullet | None:
+    stripped = line.strip()
+    if not stripped.startswith("- "):
+        return None
+
+    body = stripped[2:]
+    parsed = _parse_canonical_bullet(body)
+    if parsed is not None:
+        return parsed
+    return _parse_legacy_bullet(body)
+
+
+def _parse_canonical_bullet(body: str) -> MemoryBullet | None:
+    if len(body) < 13 or body[10:12] != ": ":
+        return None
+    try:
+        entry_date = date.fromisoformat(body[:10])
+    except ValueError:
+        return None
+    text = body[12:].strip()
+    if not text:
+        return None
+    return MemoryBullet(entry_date=entry_date, text=text)
+
+
+def _parse_legacy_bullet(body: str) -> MemoryBullet | None:
+    if len(body) < 20 or not body.startswith("("):
+        return None
+    if len(body) < 14 or body[11:14] != ") [":
+        return None
+    try:
+        entry_date = date.fromisoformat(body[1:11])
+    except ValueError:
+        return None
+
+    marker_end = body.find("] ", 14)
+    if marker_end == -1:
+        return None
+    marker = body[14:marker_end]
+    if marker not in LEGACY_MARKERS:
+        return None
+
+    text = body[marker_end + 2 :].strip()
+    if not text:
+        return None
+    return MemoryBullet(entry_date=entry_date, text=text)
+
+
+def parse_memory_document(content: str | None) -> MemoryDocument:
+    if not content:
+        return MemoryDocument()
+
+    sections: list[MemorySection] = []
+    current_heading: str | None = None
+    current_explicit = True
+    current_lines: list[MemoryLine] = []
+
+    def flush_current() -> None:
+        nonlocal current_heading, current_explicit, current_lines
+        if current_heading is None:
+            return
+        sections.append(
+            MemorySection(
+                heading=current_heading,
+                lines=current_lines,
+                explicit_heading=current_explicit,
+            )
+        )
+        current_heading = None
+        current_explicit = True
+        current_lines = []
+
+    for raw_line in content.strip().splitlines():
+        line = raw_line.rstrip()
+        if is_section_heading(line):
+            flush_current()
+            current_heading = heading_text(line)
+            current_explicit = True
+            current_lines = []
+            continue
+
+        bullet = parse_bullet_line(line)
+        if current_heading is None:
+            if bullet is None:
+                continue
+            current_heading = DEFAULT_LEGACY_SECTION
+            current_explicit = False
+            current_lines = [bullet]
+            continue
+
+        current_lines.append(bullet if bullet is not None else MemoryRawLine(text=line))
+
+    flush_current()
+    return MemoryDocument(sections=sections)
+
+
+def render_memory_document(document: MemoryDocument) -> str:
+    rendered_sections: list[str] = []
+    for section in document.sections:
+        section_lines = [f"## {section.heading}"]
+        for line in section.lines:
+            if isinstance(line, MemoryBullet):
+                section_lines.append(f"- {line.entry_date.isoformat()}: {line.text}")
+            else:
+                section_lines.append(line.text)
+        rendered_sections.append("\n".join(section_lines).strip())
+    return "\n\n".join(section for section in rendered_sections if section).strip()
+
+
+def extract_headings(memory: str | None) -> set[str]:
+    document = parse_memory_document(memory)
+    return {
+        normalize_heading(section.heading)
+        for section in document.sections
+        if section.explicit_heading
+    }
+
+
+def has_explicit_heading(content: str) -> bool:
+    return parse_memory_document(content).has_explicit_heading
+
+
+def nonstandard_bullets(content: str) -> list[str]:
+    warnings: list[str] = []
+    for line in content.splitlines():
+        stripped = line.strip()
+        if not stripped.startswith("- "):
+            continue
+        if parse_bullet_line(stripped) is not None:
+            continue
+        short = stripped[:80] + ("..." if len(stripped) > 80 else "")
+        warnings.append(f"Non-standard memory bullet: {short}")
+    return warnings

surfsense_backend/app/services/memory/prompts.py

@@ -0,0 +1,20 @@
+"""Prompts used by the memory service."""
+
+FORCED_REWRITE_PROMPT = """\
+You are a memory curator. The following memory document exceeds the character \
+limit and must be shortened.
+
+RULES:
+1. Rewrite the document to be under {target} characters.
+2. Output Markdown only. Use clear `##` headings and concise bullet points.
+3. New-format bullets should look like: `- YYYY-MM-DD: memory text`.
+4. If the input contains legacy markers like `(YYYY-MM-DD) [fact]`, preserve the
+   information but remove the inline marker in the output.
+5. Preserve durable instructions and preferences before generic facts when
+   compressing personal memory.
+6. Preserve existing headings when useful; merge duplicate headings and bullets.
+7. Output ONLY the consolidated markdown — no explanations, no wrapping.
+
+<memory_document>
+{content}
+</memory_document>"""

surfsense_backend/app/services/memory/rewrite.py

@@ -0,0 +1,35 @@
+"""LLM-backed memory rewrite helpers."""
+
+from __future__ import annotations
+
+import logging
+from typing import Any
+
+from langchain_core.messages import HumanMessage
+
+from app.services.memory.prompts import FORCED_REWRITE_PROMPT
+from app.services.memory.validation import MEMORY_HARD_LIMIT
+from app.utils.content_utils import extract_text_content
+
+logger = logging.getLogger(__name__)
+
+
+async def forced_rewrite(content: str, llm: Any) -> str | None:
+    """Use a focused LLM call to compress memory under the hard limit."""
+    try:
+        prompt = FORCED_REWRITE_PROMPT.format(
+            target=MEMORY_HARD_LIMIT,
+            content=content,
+        )
+        response = await llm.ainvoke(
+            [HumanMessage(content=prompt)],
+            config={"tags": ["surfsense:internal", "memory-rewrite"]},
+        )
+        text = extract_text_content(response.content).strip()
+        if not text:
+            logger.warning("Forced memory rewrite returned empty text")
+            return None
+        return text
+    except Exception:
+        logger.exception("Forced memory rewrite LLM call failed")
+        return None

surfsense_backend/app/services/memory/schemas.py

@@ -0,0 +1,19 @@
+"""Schemas for memory API responses and structured extraction."""
+
+from __future__ import annotations
+
+from pydantic import BaseModel
+
+
+class MemoryLimits(BaseModel):
+    """Canonical memory size limits exposed to clients."""
+
+    soft: int
+    hard: int
+
+
+class MemoryRead(BaseModel):
+    """Memory document payload returned by user and team memory APIs."""
+
+    memory_md: str
+    limits: MemoryLimits

surfsense_backend/app/services/memory/service.py

@@ -0,0 +1,247 @@
+"""Canonical read/write/reset/extract service for markdown memory."""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass, field
+from enum import StrEnum
+from typing import Any, Literal
+from uuid import UUID
+
+from sqlalchemy import select
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from app.db import SearchSpace, User
+from app.services.memory.document import parse_memory_document, render_memory_document
+from app.services.memory.rewrite import forced_rewrite
+from app.services.memory.schemas import MemoryLimits
+from app.services.memory.validation import (
+    MEMORY_HARD_LIMIT,
+    MEMORY_SOFT_LIMIT,
+    soft_limit_warning,
+    strip_preamble_to_first_heading,
+    validate_bullet_format,
+    validate_diff,
+    validate_heading_sanity,
+    validate_memory_scope,
+    validate_memory_size,
+)
+
+logger = logging.getLogger(__name__)
+
+_NO_UPDATE_SENTINELS = frozenset(
+    {
+        "NO_UPDATE",
+        "NO UPDATE",
+        "NO_CHANGE",
+        "NO CHANGE",
+    }
+)
+
+
+class MemoryScope(StrEnum):
+    USER = "user"
+    TEAM = "team"
+
+
+@dataclass(frozen=True)
+class SaveResult:
+    status: Literal["saved", "error", "no_op"]
+    message: str
+    memory_md: str = ""
+    warnings: list[str] = field(default_factory=list)
+    diff_warnings: list[str] = field(default_factory=list)
+    format_warnings: list[str] = field(default_factory=list)
+    notice: str | None = None
+
+    def to_dict(self) -> dict[str, Any]:
+        data: dict[str, Any] = {
+            "status": self.status,
+            "message": self.message,
+            "memory_md": self.memory_md,
+        }
+        if self.notice:
+            data["notice"] = self.notice
+        if self.warnings:
+            data["warnings"] = self.warnings
+            if len(self.warnings) == 1:
+                data["warning"] = self.warnings[0]
+        if self.diff_warnings:
+            data["diff_warnings"] = self.diff_warnings
+        if self.format_warnings:
+            data["format_warnings"] = self.format_warnings
+        return data
+
+
+def memory_limits() -> MemoryLimits:
+    return MemoryLimits(soft=MEMORY_SOFT_LIMIT, hard=MEMORY_HARD_LIMIT)
+
+
+def _normalize_scope(scope: MemoryScope | str) -> MemoryScope:
+    return scope if isinstance(scope, MemoryScope) else MemoryScope(scope)
+
+
+def _normalize_user_id(target_id: str | UUID) -> UUID:
+    return UUID(target_id) if isinstance(target_id, str) else target_id
+
+
+async def _load_target(
+    *,
+    scope: MemoryScope | str,
+    target_id: str | int | UUID,
+    session: AsyncSession,
+) -> User | SearchSpace | None:
+    normalized = _normalize_scope(scope)
+    if normalized is MemoryScope.USER:
+        result = await session.execute(
+            select(User).where(User.id == _normalize_user_id(target_id))  # type: ignore[arg-type]
+        )
+        return result.scalars().first()
+    result = await session.execute(
+        select(SearchSpace).where(SearchSpace.id == int(target_id))
+    )
+    return result.scalars().first()
+
+
+def _get_memory(target: User | SearchSpace, scope: MemoryScope) -> str:
+    if scope is MemoryScope.USER:
+        return getattr(target, "memory_md", None) or ""
+    return getattr(target, "shared_memory_md", None) or ""
+
+
+def _set_memory(target: User | SearchSpace, scope: MemoryScope, content: str) -> None:
+    if scope is MemoryScope.USER:
+        target.memory_md = content
+    else:
+        target.shared_memory_md = content
+
+
+async def read_memory(
+    *,
+    scope: MemoryScope | str,
+    target_id: str | int | UUID,
+    session: AsyncSession,
+) -> str:
+    normalized = _normalize_scope(scope)
+    target = await _load_target(scope=normalized, target_id=target_id, session=session)
+    if target is None:
+        return ""
+    return _get_memory(target, normalized)
+
+
+async def save_memory(
+    *,
+    scope: MemoryScope | str,
+    target_id: str | int | UUID,
+    content: str,
+    session: AsyncSession,
+    llm: Any | None = None,
+) -> SaveResult:
+    normalized = _normalize_scope(scope)
+    if not isinstance(content, str):
+        return SaveResult(
+            status="error",
+            message="Internal error: memory payload must be a string.",
+        )
+
+    target = await _load_target(scope=normalized, target_id=target_id, session=session)
+    if target is None:
+        return SaveResult(
+            status="error",
+            message="User not found."
+            if normalized is MemoryScope.USER
+            else "Search space not found.",
+        )
+
+    old_memory = _get_memory(target, normalized)
+    next_content = strip_preamble_to_first_heading(content.strip())
+    notice: str | None = None
+    warnings: list[str] = []
+
+    if next_content.upper() in _NO_UPDATE_SENTINELS:
+        return SaveResult(
+            status="no_op",
+            message="No memory update requested.",
+            memory_md=old_memory,
+        )
+
+    if len(next_content) > MEMORY_HARD_LIMIT and llm is not None:
+        rewritten = await forced_rewrite(next_content, llm)
+        if rewritten is not None and len(rewritten) < len(next_content):
+            next_content = strip_preamble_to_first_heading(rewritten)
+            notice = "Memory was automatically rewritten to fit within limits."
+
+    for validation in (
+        validate_memory_size(next_content),
+        validate_heading_sanity(next_content),
+    ):
+        if validation:
+            return SaveResult(
+                status="error",
+                message=validation["message"],
+                memory_md=old_memory,
+            )
+
+    scope_error, scope_warnings = validate_memory_scope(
+        next_content,
+        normalized.value,
+        old_memory=old_memory,
+    )
+    warnings.extend(scope_warnings)
+    if scope_error:
+        return SaveResult(
+            status="error",
+            message=scope_error["message"],
+            memory_md=old_memory,
+            warnings=warnings,
+        )
+
+    next_content = render_memory_document(parse_memory_document(next_content))
+
+    try:
+        _set_memory(target, normalized, next_content)
+        session.add(target)
+        await session.commit()
+    except Exception as e:
+        logger.exception("Failed to update %s memory: %s", normalized.value, e)
+        await session.rollback()
+        return SaveResult(
+            status="error",
+            message=f"Failed to update {normalized.value} memory: {e}",
+            memory_md=old_memory,
+        )
+
+    diff_warnings = validate_diff(old_memory, next_content)
+    format_warnings = validate_bullet_format(next_content)
+    warning = soft_limit_warning(next_content)
+    if warning:
+        warnings.append(warning)
+
+    return SaveResult(
+        status="saved",
+        message=(
+            "Memory updated."
+            if normalized is MemoryScope.USER
+            else "Team memory updated."
+        ),
+        memory_md=next_content,
+        warnings=warnings,
+        diff_warnings=diff_warnings,
+        format_warnings=format_warnings,
+        notice=notice,
+    )
+
+
+async def reset_memory(
+    *,
+    scope: MemoryScope | str,
+    target_id: str | int | UUID,
+    session: AsyncSession,
+) -> SaveResult:
+    return await save_memory(
+        scope=scope,
+        target_id=target_id,
+        content="",
+        session=session,
+        llm=None,
+    )

surfsense_backend/app/services/memory/validation.py

@@ -0,0 +1,140 @@
+"""Validation helpers for markdown-backed memory."""
+
+from __future__ import annotations
+
+from typing import Literal
+
+from app.services.memory.document import (
+    extract_headings,
+    has_explicit_heading,
+    nonstandard_bullets,
+    parse_memory_document,
+)
+
+MEMORY_SOFT_LIMIT = 18_000
+MEMORY_HARD_LIMIT = 25_000
+
+_FORBIDDEN_TEAM_HEADINGS = {
+    "preferences",
+    "instructions",
+    "personal notes",
+    "personal instructions",
+}
+
+
+def has_markdown_heading(content: str) -> bool:
+    return has_explicit_heading(content)
+
+
+def strip_preamble_to_first_heading(content: str) -> str:
+    """Drop model preamble before the first ``##`` heading, if one exists."""
+    lines = content.splitlines()
+    for index, line in enumerate(lines):
+        if line.startswith("## ") and line[3:].strip():
+            return "\n".join(lines[index:]).strip()
+    return content.strip()
+
+
+def validate_memory_size(content: str) -> dict[str, str] | None:
+    length = len(content)
+    if length > MEMORY_HARD_LIMIT:
+        return {
+            "status": "error",
+            "message": (
+                f"Memory exceeds {MEMORY_HARD_LIMIT:,} character limit "
+                f"({length:,} chars). Consolidate by merging related items, "
+                "removing outdated entries, and shortening descriptions."
+            ),
+        }
+    return None
+
+
+def validate_heading_sanity(content: str) -> dict[str, str] | None:
+    """Block long prose blobs without headings unless they are legacy bullets."""
+    stripped = content.strip()
+    if not stripped:
+        return None
+    if has_markdown_heading(stripped):
+        return None
+    if len(stripped) <= 40:
+        return None
+    if parse_memory_document(stripped).sections:
+        return None
+    return {
+        "status": "error",
+        "message": "Memory must be markdown with at least one ## heading.",
+    }
+
+
+def validate_memory_scope(
+    content: str,
+    scope: Literal["user", "team"],
+    *,
+    old_memory: str | None = None,
+) -> tuple[dict[str, str] | None, list[str]]:
+    """Reject new personal headings in team memory, grandfather existing ones."""
+    if scope != "team":
+        return None, []
+
+    old_forbidden = extract_headings(old_memory) & _FORBIDDEN_TEAM_HEADINGS
+    new_forbidden = extract_headings(content) & _FORBIDDEN_TEAM_HEADINGS
+    introduced = sorted(new_forbidden - old_forbidden)
+    grandfathered = sorted(new_forbidden & old_forbidden)
+
+    warnings: list[str] = []
+    if grandfathered:
+        warnings.append(
+            "Team memory contains legacy personal headings: "
+            + ", ".join(grandfathered)
+            + ". Please consolidate them into team-safe headings."
+        )
+    if introduced:
+        return (
+            {
+                "status": "error",
+                "message": (
+                    "Team memory cannot introduce personal headings: "
+                    + ", ".join(introduced)
+                    + ". Use team-safe headings instead."
+                ),
+            },
+            warnings,
+        )
+    return None, warnings
+
+
+def validate_bullet_format(content: str) -> list[str]:
+    return nonstandard_bullets(content)
+
+
+def validate_diff(old_memory: str | None, new_memory: str) -> list[str]:
+    if not old_memory:
+        return []
+
+    warnings: list[str] = []
+    old_headings = extract_headings(old_memory)
+    new_headings = extract_headings(new_memory)
+    dropped = old_headings - new_headings
+    if dropped:
+        names = ", ".join(sorted(dropped))
+        warnings.append(
+            f"Sections removed: {names}. If unintentional, restore them from the memory document."
+        )
+
+    old_len = len(old_memory)
+    new_len = len(new_memory)
+    if old_len > 0 and new_len < old_len * 0.4:
+        warnings.append(
+            f"Memory shrank significantly ({old_len:,} -> {new_len:,} chars). Possible data loss."
+        )
+    return warnings
+
+
+def soft_limit_warning(content: str) -> str | None:
+    length = len(content)
+    if length > MEMORY_SOFT_LIMIT:
+        return (
+            f"Memory is at {length:,}/{MEMORY_HARD_LIMIT:,} characters. "
+            "Consolidate by merging related items and removing less important entries."
+        )
+    return None

surfsense_backend/app/services/onedrive/kb_sync_service.py

@@ -1,3 +1,4 @@
+import asyncio
 import logging
 from datetime import datetime
 
@@ -95,7 +96,9 @@ class OneDriveKBSyncService:
             else:
                 logger.warning("No LLM configured — using fallback summary")
                 summary_content = f"OneDrive File: {file_name}\n\n{indexable_content}"
-                summary_embedding = embed_text(summary_content)
+                summary_embedding = await asyncio.to_thread(
+                    embed_text, summary_content
+                )
 
             chunks = await create_document_chunks(indexable_content)
             now_str = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

surfsense_backend/app/services/revert_service.py

@@ -29,6 +29,7 @@ same trap waiting to happen).
 
 from __future__ import annotations
 
+import asyncio
 import logging
 from dataclasses import dataclass
 from datetime import UTC, datetime
@@ -234,7 +235,7 @@ async def _restore_in_place_document(
             if isinstance(c, dict) and isinstance(c.get("content"), str)
         ]
         if chunk_texts:
-            chunk_embeddings = embed_texts(chunk_texts)
+            chunk_embeddings = await asyncio.to_thread(embed_texts, chunk_texts)
             session.add_all(
                 [
                     Chunk(document_id=doc.id, content=text, embedding=embedding)
@@ -244,7 +245,9 @@ async def _restore_in_place_document(
                 ]
             )
             if isinstance(revision.content_before, str):
-                doc.embedding = embed_texts([revision.content_before])[0]
+                doc.embedding = (
+                    await asyncio.to_thread(embed_texts, [revision.content_before])
+                )[0]
 
     doc.updated_at = datetime.now(UTC)
     return RevertOutcome(status="ok", message="Document restored from snapshot.")
@@ -320,7 +323,7 @@ async def _reinsert_document_from_revision(
     session.add(new_doc)
     await session.flush()
 
-    new_doc.embedding = embed_texts([content])[0]
+    new_doc.embedding = (await asyncio.to_thread(embed_texts, [content]))[0]
     chunk_texts = []
     chunks_before = revision.chunks_before
     if isinstance(chunks_before, list):
@@ -330,7 +333,7 @@ async def _reinsert_document_from_revision(
             if isinstance(c, dict) and isinstance(c.get("content"), str)
         ]
     if chunk_texts:
-        chunk_embeddings = embed_texts(chunk_texts)
+        chunk_embeddings = await asyncio.to_thread(embed_texts, chunk_texts)
         session.add_all(
             [
                 Chunk(document_id=new_doc.id, content=text, embedding=embedding)

surfsense_backend/app/services/token_tracking_service.py

@@ -325,6 +325,24 @@ class TokenTrackingCallback(CustomLogger):
             total_tokens = getattr(usage, "total_tokens", 0) or 0
             call_kind = "chat"
 
+        # Prompt-cache accounting. LiteLLM normalizes every provider's cache
+        # fields onto ``usage.prompt_tokens_details``:
+        # - ``cached_tokens``         — cache reads (OpenAI/Azure native, DeepSeek
+        #                               mapped from ``prompt_cache_hit_tokens``,
+        #                               Anthropic mapped from ``cache_read_input_tokens``).
+        # - ``cache_creation_tokens`` — cache writes (Anthropic only; OpenAI/Azure
+        #                               do not expose a write count).
+        # See ``litellm.types.utils.Usage.__init__`` for the mapping.
+        cached_tokens = 0
+        cache_creation_tokens = 0
+        if not is_image:
+            prompt_details = getattr(usage, "prompt_tokens_details", None)
+            if prompt_details is not None:
+                cached_tokens = getattr(prompt_details, "cached_tokens", 0) or 0
+                cache_creation_tokens = (
+                    getattr(prompt_details, "cache_creation_tokens", 0) or 0
+                )
+
         model = kwargs.get("model", "unknown")
 
         cost_usd = _extract_cost_usd(
@@ -357,9 +375,23 @@ class TokenTrackingCallback(CustomLogger):
             cost_micros=cost_micros,
             call_kind=call_kind,
         )
+
+        # Per-LLM-call wall-clock latency (LiteLLM passes datetime objects).
+        call_latency_s: float | None = None
+        try:
+            if start_time is not None and end_time is not None:
+                delta = end_time - start_time
+                call_latency_s = getattr(delta, "total_seconds", lambda: float(delta))()
+        except Exception:
+            call_latency_s = None
+
+        cache_hit_ratio: float | None = None
+        if prompt_tokens > 0 and (cached_tokens > 0 or cache_creation_tokens > 0):
+            cache_hit_ratio = cached_tokens / prompt_tokens
+
         logger.info(
             "[TokenTracking] Captured: model=%s kind=%s prompt=%d completion=%d total=%d "
-            "cost=$%.6f (%d micros) (accumulator now has %d calls)",
+            "cost=$%.6f (%d micros) (accumulator now has %d calls)%s%s",
             model,
             call_kind,
             prompt_tokens,
@@ -368,6 +400,17 @@ class TokenTrackingCallback(CustomLogger):
             cost_usd,
             cost_micros,
             len(acc.calls),
+            f" latency={call_latency_s:.3f}s" if call_latency_s is not None else "",
+            (
+                f" cache_read={cached_tokens} cache_write={cache_creation_tokens}"
+                f" hit_ratio={cache_hit_ratio:.1%}"
+                if cache_hit_ratio is not None
+                else (
+                    f" cache_read={cached_tokens} cache_write={cache_creation_tokens}"
+                    if (cached_tokens or cache_creation_tokens)
+                    else ""
+                )
+            ),
         )
 
 

surfsense_backend/app/tasks/chat/stream_new_chat.py

@@ -39,10 +39,6 @@ from app.agents.new_chat.llm_config import (
     load_agent_config,
     load_global_llm_config_by_id,
 )
-from app.agents.new_chat.memory_extraction import (
-    extract_and_save_memory,
-    extract_and_save_team_memory,
-)
 from app.agents.new_chat.mention_resolver import resolve_mentions, substitute_in_text
 from app.agents.new_chat.middleware.busy_mutex import (
     end_turn,
@@ -64,8 +60,6 @@ from app.db import (
 )
 from app.prompts import TITLE_GENERATION_PROMPT
 from app.services.auto_model_pin_service import (
-    is_recently_healthy,
-    mark_healthy,
     mark_runtime_cooldown,
     resolve_or_get_pinned_llm_config_id,
 )
@@ -283,7 +277,6 @@ class StreamResult:
     accumulated_text: str = ""
     is_interrupted: bool = False
     sandbox_files: list[str] = field(default_factory=list)
-    agent_called_update_memory: bool = False
     request_id: str | None = None
     turn_id: str = ""
     filesystem_mode: str = "cloud"
@@ -506,54 +499,6 @@ def _is_provider_rate_limited(exc: BaseException) -> bool:
     )
 
 
-_PREFLIGHT_TIMEOUT_SEC: float = 2.5
-_PREFLIGHT_MAX_TOKENS: int = 1
-
-
-async def _preflight_llm(llm: Any) -> None:
-    """Issue a minimal completion to confirm the pinned model isn't 429'ing.
-
-    Used before agent build / planner / classifier / title-gen so a known-bad
-    free OpenRouter deployment is detected and repinned before it cascades
-    into multiple wasted internal calls. The probe is intentionally cheap:
-    one token, low timeout, tagged ``surfsense:internal`` so token tracking
-    and SSE pipelines treat it as overhead rather than user output.
-
-    Raises the original exception when the provider responds with a
-    rate-limit-shaped error so the caller can drive the cooldown/repin
-    branch via :func:`_is_provider_rate_limited`. Other transient failures
-    are swallowed — the caller continues to the normal stream path and the
-    in-stream recovery loop remains the safety net.
-    """
-    from litellm import acompletion
-
-    model = getattr(llm, "model", None)
-    if not model or model == "auto":
-        # Auto-mode router doesn't have a single deployment to ping; the
-        # router itself handles per-deployment rate-limit accounting.
-        return
-
-    try:
-        await acompletion(
-            model=model,
-            messages=[{"role": "user", "content": "ping"}],
-            api_key=getattr(llm, "api_key", None),
-            api_base=getattr(llm, "api_base", None),
-            max_tokens=_PREFLIGHT_MAX_TOKENS,
-            timeout=_PREFLIGHT_TIMEOUT_SEC,
-            stream=False,
-            metadata={"tags": ["surfsense:internal", "auto-pin-preflight"]},
-        )
-    except Exception as exc:
-        if _is_provider_rate_limited(exc):
-            raise
-        logging.getLogger(__name__).debug(
-            "auto_pin_preflight non_rate_limit_error model=%s err=%s",
-            model,
-            exc,
-        )
-
-
 async def _build_main_agent_for_thread(
     agent_factory: Any,
     *,
@@ -571,9 +516,9 @@ async def _build_main_agent_for_thread(
     disabled_tools: list[str] | None = None,
     mentioned_document_ids: list[int] | None = None,
 ) -> Any:
-    """Single (re)build path so the agent factory cannot drift across
-    initial build, preflight repin, and mid-stream 429 recovery for one
-    ``thread_id``: a graph swap mid-turn would corrupt checkpointer state."""
+    """Single (re)build path so the agent factory cannot drift across the
+    initial build and mid-stream 429 recovery for one ``thread_id``: a
+    graph swap mid-turn would corrupt checkpointer state."""
     return await agent_factory(
         llm=llm,
         search_space_id=search_space_id,
@@ -591,29 +536,6 @@ async def _build_main_agent_for_thread(
     )
 
 
-async def _settle_speculative_agent_build(task: asyncio.Task[Any]) -> None:
-    """Wait for a discarded speculative agent build to release shared state.
-
-    Used by the parallel preflight + agent-build path. The speculative build
-    closes over the request-scoped ``AsyncSession`` (for the brief connector
-    discovery / tool-factory window before its CPU work moves into a worker
-    thread). If preflight reports a 429 we want to fall back to the original
-    repin → reload → rebuild path, but we MUST NOT touch ``session`` again
-    until any in-flight session work owned by the speculative build has
-    fully settled — :class:`sqlalchemy.ext.asyncio.AsyncSession` is not
-    concurrency-safe and the same hazard cost us a hard ``InvalidRequestError``
-    earlier in this PR (see ``connector_service`` parallel-gather revert).
-
-    We simply ``await`` the task and swallow any exception: in this path the
-    build's outcome is irrelevant — success populates the agent cache (a free
-    side effect), failure is discarded. The wasted CPU is acceptable since
-    429 fallbacks are rare and the original sequential code also paid the
-    full build cost on the same path.
-    """
-    with contextlib.suppress(BaseException):
-        await task
-
-
 def _classify_stream_exception(
     exc: Exception,
     *,
@@ -1241,39 +1163,6 @@ async def stream_new_chat(
             yield streaming_service.format_done()
             return
 
-        # Auto-mode preflight ping. Runs ONLY for thread-pinned auto cfgs
-        # (negative ids selected via ``resolve_or_get_pinned_llm_config_id``)
-        # whose health hasn't already been confirmed within the TTL window.
-        # Detecting a 429 here lets us repin BEFORE the planner/classifier/
-        # title-generation LLM calls fan out and each independently hit the
-        # same upstream rate limit.
-        #
-        # PERF: preflight is a network round-trip to the LLM provider (~1-5s)
-        # and is independent of the agent build (CPU-bound, ~5-7s). They used
-        # to run sequentially → ``preflight + build`` on cold cache = 11.5s.
-        # We now kick off preflight as a background task FIRST, then run the
-        # synchronous setup work and the agent build in parallel. In the
-        # success path (the common case) total wall time drops to roughly
-        # ``max(preflight, build)`` — the preflight finishes during the
-        # agent compile and we just consume its result. In the rare 429
-        # path the speculative build is awaited to completion (so its
-        # session usage is fully released) via
-        # :func:`_settle_speculative_agent_build`, then discarded, and
-        # we fall back to the original repin-and-rebuild flow.
-        preflight_needed = (
-            requested_llm_config_id == 0
-            and llm_config_id < 0
-            and not is_recently_healthy(llm_config_id)
-        )
-        preflight_task: asyncio.Task[None] | None = None
-        _t_preflight = 0.0
-        if preflight_needed:
-            _t_preflight = time.perf_counter()
-            preflight_task = asyncio.create_task(
-                _preflight_llm(llm),
-                name=f"auto_pin_preflight:{llm_config_id}",
-            )
-
         # Create connector service
         _t0 = time.perf_counter()
         connector_service = ConnectorService(session, search_space_id=search_space_id)
@@ -1307,136 +1196,26 @@ async def stream_new_chat(
             if use_multi_agent
             else create_surfsense_deep_agent
         )
-        # Speculative agent build — runs in parallel with the preflight
-        # task (if any). Built with the *current* ``llm`` / ``agent_config``;
-        # if preflight reports 429 we will discard this future and rebuild
-        # against the freshly pinned config below.
-        agent_build_task = asyncio.create_task(
-            _build_main_agent_for_thread(
-                agent_factory,
-                llm=llm,
-                search_space_id=search_space_id,
-                db_session=session,
-                connector_service=connector_service,
-                checkpointer=checkpointer,
-                user_id=user_id,
-                thread_id=chat_id,
-                agent_config=agent_config,
-                firecrawl_api_key=firecrawl_api_key,
-                thread_visibility=visibility,
-                filesystem_selection=filesystem_selection,
-                disabled_tools=disabled_tools,
-                mentioned_document_ids=mentioned_document_ids,
-            ),
-            name="agent_build:stream_new_chat",
+        # Build the agent inline. Provider 429s surface through the
+        # in-stream recovery loop below (``_is_provider_rate_limited``),
+        # which repins the thread to an eligible alternative config and
+        # rebuilds the agent before the user sees any output.
+        agent = await _build_main_agent_for_thread(
+            agent_factory,
+            llm=llm,
+            search_space_id=search_space_id,
+            db_session=session,
+            connector_service=connector_service,
+            checkpointer=checkpointer,
+            user_id=user_id,
+            thread_id=chat_id,
+            agent_config=agent_config,
+            firecrawl_api_key=firecrawl_api_key,
+            thread_visibility=visibility,
+            filesystem_selection=filesystem_selection,
+            disabled_tools=disabled_tools,
+            mentioned_document_ids=mentioned_document_ids,
         )
-
-        agent: Any = None
-        if preflight_task is not None:
-            try:
-                await preflight_task
-                mark_healthy(llm_config_id)
-                _perf_log.info(
-                    "[stream_new_chat] auto_pin_preflight ok config_id=%s took=%.3fs (parallel)",
-                    llm_config_id,
-                    time.perf_counter() - _t_preflight,
-                )
-            except Exception as preflight_exc:
-                # Both branches below need the session: the non-429 path
-                # may unwind via cleanup that uses ``session``, and the
-                # 429 path explicitly calls ``resolve_or_get_pinned_llm_config_id``
-                # against it. Wait for the speculative build to release its
-                # session usage before we proceed.
-                await _settle_speculative_agent_build(agent_build_task)
-                if not _is_provider_rate_limited(preflight_exc):
-                    raise
-                # 429: speculative agent is discarded; run the original
-                # repin → reload → rebuild path against the freshly
-                # pinned config.
-                previous_config_id = llm_config_id
-                mark_runtime_cooldown(
-                    previous_config_id, reason="preflight_rate_limited"
-                )
-                try:
-                    llm_config_id = (
-                        await resolve_or_get_pinned_llm_config_id(
-                            session,
-                            thread_id=chat_id,
-                            search_space_id=search_space_id,
-                            user_id=user_id,
-                            selected_llm_config_id=0,
-                            exclude_config_ids={previous_config_id},
-                            requires_image_input=_requires_image_input,
-                        )
-                    ).resolved_llm_config_id
-                except ValueError as pin_error:
-                    yield _emit_stream_error(
-                        message=str(pin_error),
-                        error_kind="server_error",
-                        error_code="SERVER_ERROR",
-                    )
-                    yield streaming_service.format_done()
-                    return
-
-                llm, agent_config, llm_load_error = await _load_llm_bundle(
-                    llm_config_id
-                )
-                if llm_load_error or not llm:
-                    yield _emit_stream_error(
-                        message=llm_load_error or "Failed to create LLM instance",
-                        error_kind="server_error",
-                        error_code="SERVER_ERROR",
-                    )
-                    yield streaming_service.format_done()
-                    return
-                # Trust the freshly-resolved cfg for the remainder of this
-                # turn rather than recursing into another preflight; the
-                # in-stream 429 recovery loop is still in place as the
-                # safety net if even this fallback hits an upstream cap.
-                mark_healthy(llm_config_id)
-                _log_chat_stream_error(
-                    flow=flow,
-                    error_kind="rate_limited",
-                    error_code="RATE_LIMITED",
-                    severity="info",
-                    is_expected=True,
-                    request_id=request_id,
-                    thread_id=chat_id,
-                    search_space_id=search_space_id,
-                    user_id=user_id,
-                    message=(
-                        "Auto-pinned model failed preflight; switched to another "
-                        "eligible model and continuing."
-                    ),
-                    extra={
-                        "auto_runtime_recover": True,
-                        "preflight": True,
-                        "previous_config_id": previous_config_id,
-                        "fallback_config_id": llm_config_id,
-                    },
-                )
-                # Rebuild against the new llm/agent_config. Sequential
-                # here because we no longer have anything to overlap with.
-                agent = await agent_factory(
-                    llm=llm,
-                    search_space_id=search_space_id,
-                    db_session=session,
-                    connector_service=connector_service,
-                    checkpointer=checkpointer,
-                    user_id=user_id,
-                    thread_id=chat_id,
-                    agent_config=agent_config,
-                    firecrawl_api_key=firecrawl_api_key,
-                    thread_visibility=visibility,
-                    disabled_tools=disabled_tools,
-                    mentioned_document_ids=mentioned_document_ids,
-                    filesystem_selection=filesystem_selection,
-                )
-
-        if agent is None:
-            # Either no preflight was needed, or preflight succeeded —
-            # in both cases the speculative build is the agent we want.
-            agent = await agent_build_task
         _perf_log.info(
             "[stream_new_chat] Agent created in %.3fs", time.perf_counter() - _t0
         )
@@ -2208,36 +1987,6 @@ async def stream_new_chat(
                 },
             )
 
-        # Fire background memory extraction if the agent didn't handle it.
-        # Shared threads write to team memory; private threads write to user memory.
-        if not stream_result.agent_called_update_memory:
-            memory_seed = user_query.strip() or (
-                f"[{len(user_image_data_urls or [])} image(s)]"
-                if user_image_data_urls
-                else "(message)"
-            )
-            if visibility == ChatVisibility.SEARCH_SPACE:
-                task = asyncio.create_task(
-                    extract_and_save_team_memory(
-                        user_message=memory_seed,
-                        search_space_id=search_space_id,
-                        llm=llm,
-                        author_display_name=current_user_display_name,
-                    )
-                )
-                _background_tasks.add(task)
-                task.add_done_callback(_background_tasks.discard)
-            elif user_id:
-                task = asyncio.create_task(
-                    extract_and_save_memory(
-                        user_message=memory_seed,
-                        user_id=user_id,
-                        llm=llm,
-                    )
-                )
-                _background_tasks.add(task)
-                task.add_done_callback(_background_tasks.discard)
-
         # Finish the step and message
         yield streaming_service.format_data("turn-status", {"status": "idle"})
         yield streaming_service.format_finish_step()
@@ -2682,25 +2431,6 @@ async def stream_resume_chat(
             yield streaming_service.format_done()
             return
 
-        # Auto-mode preflight ping (resume path). Mirrors ``stream_new_chat``:
-        # one cheap probe before the agent is rebuilt so a 429'd pin gets
-        # repinned without burning planner/classifier/title calls first.
-        # See ``stream_new_chat`` for the full rationale on the speculative
-        # parallel build pattern below.
-        preflight_needed = (
-            requested_llm_config_id == 0
-            and llm_config_id < 0
-            and not is_recently_healthy(llm_config_id)
-        )
-        preflight_task: asyncio.Task[None] | None = None
-        _t_preflight = 0.0
-        if preflight_needed:
-            _t_preflight = time.perf_counter()
-            preflight_task = asyncio.create_task(
-                _preflight_llm(llm),
-                name=f"auto_pin_preflight_resume:{llm_config_id}",
-            )
-
         _t0 = time.perf_counter()
         connector_service = ConnectorService(session, search_space_id=search_space_id)
 
@@ -2730,115 +2460,25 @@ async def stream_resume_chat(
             if _app_config.MULTI_AGENT_CHAT_ENABLED
             else create_surfsense_deep_agent
         )
-        agent_build_task = asyncio.create_task(
-            _build_main_agent_for_thread(
-                agent_factory,
-                llm=llm,
-                search_space_id=search_space_id,
-                db_session=session,
-                connector_service=connector_service,
-                checkpointer=checkpointer,
-                user_id=user_id,
-                thread_id=chat_id,
-                agent_config=agent_config,
-                firecrawl_api_key=firecrawl_api_key,
-                thread_visibility=visibility,
-                filesystem_selection=filesystem_selection,
-                disabled_tools=disabled_tools,
-            ),
-            name="agent_build:stream_resume",
+        # Build the agent inline. Provider 429s are handled by the
+        # in-stream recovery loop, which repins to an eligible
+        # alternative config and rebuilds the agent before the user sees
+        # any output.
+        agent = await _build_main_agent_for_thread(
+            agent_factory,
+            llm=llm,
+            search_space_id=search_space_id,
+            db_session=session,
+            connector_service=connector_service,
+            checkpointer=checkpointer,
+            user_id=user_id,
+            thread_id=chat_id,
+            agent_config=agent_config,
+            firecrawl_api_key=firecrawl_api_key,
+            thread_visibility=visibility,
+            filesystem_selection=filesystem_selection,
+            disabled_tools=disabled_tools,
         )
-
-        agent: Any = None
-        if preflight_task is not None:
-            try:
-                await preflight_task
-                mark_healthy(llm_config_id)
-                _perf_log.info(
-                    "[stream_resume] auto_pin_preflight ok config_id=%s took=%.3fs (parallel)",
-                    llm_config_id,
-                    time.perf_counter() - _t_preflight,
-                )
-            except Exception as preflight_exc:
-                # Same session-safety rationale as ``stream_new_chat``.
-                await _settle_speculative_agent_build(agent_build_task)
-                if not _is_provider_rate_limited(preflight_exc):
-                    raise
-                previous_config_id = llm_config_id
-                mark_runtime_cooldown(
-                    previous_config_id, reason="preflight_rate_limited"
-                )
-                try:
-                    llm_config_id = (
-                        await resolve_or_get_pinned_llm_config_id(
-                            session,
-                            thread_id=chat_id,
-                            search_space_id=search_space_id,
-                            user_id=user_id,
-                            selected_llm_config_id=0,
-                            exclude_config_ids={previous_config_id},
-                        )
-                    ).resolved_llm_config_id
-                except ValueError as pin_error:
-                    yield _emit_stream_error(
-                        message=str(pin_error),
-                        error_kind="server_error",
-                        error_code="SERVER_ERROR",
-                    )
-                    yield streaming_service.format_done()
-                    return
-
-                llm, agent_config, llm_load_error = await _load_llm_bundle(
-                    llm_config_id
-                )
-                if llm_load_error or not llm:
-                    yield _emit_stream_error(
-                        message=llm_load_error or "Failed to create LLM instance",
-                        error_kind="server_error",
-                        error_code="SERVER_ERROR",
-                    )
-                    yield streaming_service.format_done()
-                    return
-                mark_healthy(llm_config_id)
-                _log_chat_stream_error(
-                    flow="resume",
-                    error_kind="rate_limited",
-                    error_code="RATE_LIMITED",
-                    severity="info",
-                    is_expected=True,
-                    request_id=request_id,
-                    thread_id=chat_id,
-                    search_space_id=search_space_id,
-                    user_id=user_id,
-                    message=(
-                        "Auto-pinned model failed preflight; switched to another "
-                        "eligible model and continuing."
-                    ),
-                    extra={
-                        "auto_runtime_recover": True,
-                        "preflight": True,
-                        "previous_config_id": previous_config_id,
-                        "fallback_config_id": llm_config_id,
-                    },
-                )
-                agent = await _build_main_agent_for_thread(
-                    agent_factory,
-                    llm=llm,
-                    search_space_id=search_space_id,
-                    db_session=session,
-                    connector_service=connector_service,
-                    checkpointer=checkpointer,
-                    user_id=user_id,
-                    thread_id=chat_id,
-                    agent_config=agent_config,
-                    firecrawl_api_key=firecrawl_api_key,
-                    thread_visibility=visibility,
-                    filesystem_selection=filesystem_selection,
-                    disabled_tools=disabled_tools,
-                )
-
-        if agent is None:
-            agent = await agent_build_task
         _perf_log.info(
             "[stream_resume] Agent created in %.3fs", time.perf_counter() - _t0
         )

surfsense_backend/app/tasks/chat/streaming/graph_stream/event_stream.py

@@ -48,4 +48,3 @@ async def stream_output(
         yield frame
 
     result.accumulated_text = state.accumulated_text
-    result.agent_called_update_memory = state.called_update_memory

surfsense_backend/app/tasks/chat/streaming/graph_stream/result.py

@@ -11,7 +11,6 @@ class StreamingResult:
     accumulated_text: str = ""
     is_interrupted: bool = False
     sandbox_files: list[str] = field(default_factory=list)
-    agent_called_update_memory: bool = False
     request_id: str | None = None
     turn_id: str = ""
     filesystem_mode: str = "cloud"

surfsense_backend/app/tasks/chat/streaming/handlers/tool_end.py

@@ -36,9 +36,6 @@ def iter_tool_end_frames(
     raw_output = event.get("data", {}).get("output", "")
     staged_file_path = state.file_path_by_run.pop(run_id, None) if run_id else None
 
-    if tool_name == "update_memory":
-        state.called_update_memory = True
-
     if hasattr(raw_output, "content"):
         content = raw_output.content
         if isinstance(content, str):

surfsense_backend/app/tasks/chat/streaming/relay/state.py

@@ -32,7 +32,6 @@ class AgentEventRelayState:
     last_active_step_items: list[str] = field(default_factory=list)
     just_finished_tool: bool = False
     active_tool_depth: int = 0
-    called_update_memory: bool = False
     current_reasoning_id: str | None = None
     pending_tool_call_chunks: list[dict[str, Any]] = field(default_factory=list)
     lc_tool_call_id_by_run: dict[str, str] = field(default_factory=dict)

surfsense_backend/app/tasks/connector_indexers/discord_indexer.py

@@ -670,7 +670,9 @@ async def index_discord_messages(
 
                 # Heavy processing (embeddings, chunks)
                 chunks = await create_document_chunks(item["combined_document_string"])
-                doc_embedding = embed_text(item["combined_document_string"])
+                doc_embedding = await asyncio.to_thread(
+                    embed_text, item["combined_document_string"]
+                )
 
                 # Update document to READY with actual content
                 document.title = f"{item['guild_name']}#{item['channel_name']}"

surfsense_backend/app/tasks/connector_indexers/luma_indexer.py

@@ -6,6 +6,7 @@ Implements 2-phase document status updates for real-time UI feedback:
 - Phase 2: Process each event: pending → processing → ready/failed
 """
 
+import asyncio
 import time
 from collections.abc import Awaitable, Callable
 from datetime import datetime, timedelta
@@ -465,7 +466,9 @@ async def index_luma_events(
                     summary_content = (
                         f"Luma Event: {item['event_name']}\n\n{item['event_markdown']}"
                     )
-                    summary_embedding = embed_text(summary_content)
+                    summary_embedding = await asyncio.to_thread(
+                        embed_text, summary_content
+                    )
 
                 chunks = await create_document_chunks(item["event_markdown"])
 

surfsense_backend/app/tasks/connector_indexers/teams_indexer.py

@@ -9,6 +9,7 @@ Uses 2-phase document status updates for real-time UI feedback:
 - Phase 2: Process each document: pending → processing → ready/failed
 """
 
+import asyncio
 import time
 from collections.abc import Awaitable, Callable
 from datetime import UTC, datetime
@@ -581,7 +582,9 @@ async def index_teams_messages(
 
                 # Heavy processing (embeddings, chunks)
                 chunks = await create_document_chunks(item["combined_document_string"])
-                doc_embedding = embed_text(item["combined_document_string"])
+                doc_embedding = await asyncio.to_thread(
+                    embed_text, item["combined_document_string"]
+                )
 
                 # Update document to READY with actual content
                 document.title = f"{item['team_name']} - {item['channel_name']}"

surfsense_backend/app/tasks/document_processors/_save.py

@@ -2,6 +2,7 @@
 Unified document save/update logic for file processors.
 """
 
+import asyncio
 import logging
 
 from sqlalchemy.exc import SQLAlchemyError
@@ -43,7 +44,7 @@ async def _generate_summary(
     """
     if not enable_summary:
         summary = f"File: {file_name}\n\n{markdown_content[:4000]}"
-        return summary, embed_text(summary)
+        return summary, await asyncio.to_thread(embed_text, summary)
 
     if etl_service == "DOCLING":
         from app.services.docling_service import create_docling_service
@@ -65,7 +66,7 @@ async def _generate_summary(
                 parts.append(f"**{formatted_key}:** {value}")
 
         enhanced = "\n".join(parts) + "\n\n# DOCUMENT SUMMARY\n\n" + summary_text
-        return enhanced, embed_text(enhanced)
+        return enhanced, await asyncio.to_thread(embed_text, enhanced)
 
     # Standard summary (Unstructured / LlamaCloud / others)
     meta = {

surfsense_backend/app/utils/document_converters.py

@@ -1,3 +1,4 @@
+import asyncio
 import hashlib
 import logging
 import threading
@@ -221,7 +222,9 @@ async def generate_document_summary(
     else:
         enhanced_summary_content = summary_content
 
-    summary_embedding = embed_text(enhanced_summary_content)
+    summary_embedding = await asyncio.to_thread(
+        embed_text, enhanced_summary_content
+    )
 
     return enhanced_summary_content, summary_embedding
 
@@ -237,7 +240,7 @@ async def create_document_chunks(content: str) -> list[Chunk]:
         List of Chunk objects with embeddings
     """
     chunk_texts = [c.text for c in config.chunker_instance.chunk(content)]
-    chunk_embeddings = embed_texts(chunk_texts)
+    chunk_embeddings = await asyncio.to_thread(embed_texts, chunk_texts)
     return [
         Chunk(content=text, embedding=emb)
         for text, emb in zip(chunk_texts, chunk_embeddings, strict=False)

surfsense_backend/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "surf-new-backend"
-version = "0.0.24"
+version = "0.0.25"
 description = "SurfSense Backend"
 requires-python = ">=3.12"
 dependencies = [

surfsense_backend/tests/unit/agents/new_chat/test_memory_response_content.py

@@ -2,28 +2,12 @@
 
 import pytest
 
-from app.agents.new_chat.tools.update_memory import _save_memory
+from app.services.memory import MemoryScope, save_memory
 from app.utils.content_utils import extract_text_content
 
 pytestmark = pytest.mark.unit
 
 
-class _Recorder:
-    def __init__(self) -> None:
-        self.applied_content: str | None = None
-        self.commit_calls = 0
-        self.rollback_calls = 0
-
-    def apply(self, content: str) -> None:
-        self.applied_content = content
-
-    async def commit(self) -> None:
-        self.commit_calls += 1
-
-    async def rollback(self) -> None:
-        self.rollback_calls += 1
-
-
 def test_extract_text_content_keeps_no_update_bare_string_from_content_blocks() -> None:
     content = [
         {"type": "thinking", "thinking": "No"},
@@ -69,21 +53,12 @@ def test_extract_text_content_preserves_plain_string_responses() -> None:
 
 @pytest.mark.asyncio
 async def test_save_memory_rejects_non_string_payload_before_commit() -> None:
-    recorder = _Recorder()
-
-    result = await _save_memory(
-        updated_memory=["NO_UPDATE"],  # type: ignore[arg-type]
-        old_memory=None,
-        llm=None,
-        apply_fn=recorder.apply,
-        commit_fn=recorder.commit,
-        rollback_fn=recorder.rollback,
-        label="memory",
-        scope="user",
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content=["NO_UPDATE"],  # type: ignore[arg-type]
+        session=None,  # type: ignore[arg-type]
     )
 
-    assert result["status"] == "error"
-    assert "must be a string" in result["message"]
-    assert recorder.applied_content is None
-    assert recorder.commit_calls == 0
-    assert recorder.rollback_calls == 0
+    assert result.status == "error"
+    assert "must be a string" in result.message

surfsense_backend/tests/unit/agents/new_chat/test_prompt_caching.py

@@ -12,13 +12,19 @@ prompt caching. It mutates ``llm.model_kwargs`` so the kwargs flow to
    the deepagent stack accumulates multiple ``SystemMessage``\ s in
    ``state["messages"]`` and ``role: system`` would tag every one of
    them, blowing past Anthropic's 4-block ``cache_control`` cap.
-2. Adds ``prompt_cache_key``/``prompt_cache_retention`` only for
-   single-model OPENAI/DEEPSEEK/XAI configs (where OpenAI's automatic
-   prompt-cache surface is available).
-3. Treats ``ChatLiteLLMRouter`` (auto-mode) as universal-only — no
-   OpenAI-only kwargs because the router fans out across providers.
-4. Idempotent: user-supplied values in ``model_kwargs`` are preserved.
-5. Defensive: LLMs without a writable ``model_kwargs`` are silently
+2. Adds ``prompt_cache_key`` for OPENAI/DEEPSEEK/XAI/AZURE/AZURE_OPENAI
+   configs (Microsoft's Azure transformer was added to LiteLLM in
+   https://github.com/BerriAI/litellm/pull/20989, Feb 2026).
+3. Adds ``prompt_cache_retention="24h"`` ONLY for OPENAI/DEEPSEEK/XAI.
+   Azure's server-side support landed in Microsoft's docs on 2026-05-13
+   but LiteLLM 1.83.14 hasn't wired it through yet, so we let Azure use
+   its default in-memory retention rather than send a param that
+   ``litellm.drop_params`` would silently strip.
+4. Treats ``ChatLiteLLMRouter`` (auto-mode) as universal-only — no
+   destination-specific kwargs because the router fans out across
+   providers.
+5. Idempotent: user-supplied values in ``model_kwargs`` are preserved.
+6. Defensive: LLMs without a writable ``model_kwargs`` are silently
    skipped rather than raising.
 """
 
@@ -191,9 +197,9 @@ def test_does_not_overwrite_user_supplied_prompt_cache_key() -> None:
 
 @pytest.mark.parametrize("provider", ["OPENAI", "DEEPSEEK", "XAI"])
 def test_sets_openai_family_extras(provider: str) -> None:
-    """OpenAI-style providers gain ``prompt_cache_key`` (raises hit rate
-    via routing affinity) and ``prompt_cache_retention="24h"`` (extends
-    cache TTL beyond the default 5-10 min)."""
+    """Native OpenAI-style providers gain ``prompt_cache_key`` (raises
+    hit rate via routing affinity) and ``prompt_cache_retention="24h"``
+    (extends cache TTL beyond the default 5-10 min)."""
     cfg = _make_cfg(provider=provider)
     llm = _FakeLLM()
 
@@ -203,6 +209,27 @@ def test_sets_openai_family_extras(provider: str) -> None:
     assert llm.model_kwargs["prompt_cache_retention"] == "24h"
 
 
+@pytest.mark.parametrize("provider", ["AZURE", "AZURE_OPENAI"])
+def test_azure_gets_prompt_cache_key_only(provider: str) -> None:
+    """Azure configs gain ``prompt_cache_key`` for routing affinity
+    (Microsoft auto-caches every GPT-4o+ deployment at ≥1024 tokens;
+    the key clusters same-prefix requests on the same backend GPU pool
+    so hit rate climbs). They DO NOT get ``prompt_cache_retention``
+    because LiteLLM 1.83.14's Azure transformer omits it from its
+    supported params list — ``drop_params`` would silently strip it.
+    Azure's default in-memory retention (5-10 min, max 1 h) is already
+    enough to cover intra-conversation turns; revisit when LiteLLM
+    bumps Azure to match its OpenAI surface."""
+    cfg = _make_cfg(provider=provider, model_name="gpt-5.4")
+    llm = _FakeLLM(model="azure/gpt-5.4")
+
+    apply_litellm_prompt_caching(llm, agent_config=cfg, thread_id=42)
+
+    assert llm.model_kwargs["prompt_cache_key"] == "surfsense-thread-42"
+    assert "prompt_cache_retention" not in llm.model_kwargs
+    assert "cache_control_injection_points" in llm.model_kwargs
+
+
 def test_skips_prompt_cache_key_when_no_thread_id() -> None:
     """Without a thread id we can't construct a per-thread key. Retention
     is still useful so we set it (it's free)."""
@@ -215,12 +242,26 @@ def test_skips_prompt_cache_key_when_no_thread_id() -> None:
     assert llm.model_kwargs["prompt_cache_retention"] == "24h"
 
 
+def test_azure_skips_prompt_cache_key_when_no_thread_id() -> None:
+    """Azure without a thread id ends up with no extras (retention is
+    Azure-skipped, key needs a thread id) — universal injection points
+    still land."""
+    cfg = _make_cfg(provider="AZURE", model_name="gpt-5.4")
+    llm = _FakeLLM(model="azure/gpt-5.4")
+
+    apply_litellm_prompt_caching(llm, agent_config=cfg, thread_id=None)
+
+    assert "prompt_cache_key" not in llm.model_kwargs
+    assert "prompt_cache_retention" not in llm.model_kwargs
+    assert "cache_control_injection_points" in llm.model_kwargs
+
+
 @pytest.mark.parametrize(
     "provider",
     ["ANTHROPIC", "BEDROCK", "VERTEX_AI", "GOOGLE_AI_STUDIO", "GROQ", "MOONSHOT"],
 )
 def test_no_openai_extras_for_other_providers(provider: str) -> None:
-    """Non-OpenAI-family providers don't expose ``prompt_cache_key`` —
+    """Non-OpenAI-style providers don't expose ``prompt_cache_key`` —
     skip it. ``cache_control_injection_points`` is still set (universal)."""
     cfg = _make_cfg(provider=provider)
     llm = _FakeLLM()

surfsense_backend/tests/unit/agents/new_chat/tools/test_mcp_tools_cache.py

@@ -0,0 +1,130 @@
+"""Unit tests for ``mcp_tools_cache``."""
+
+from __future__ import annotations
+
+from datetime import UTC, datetime
+from types import SimpleNamespace
+
+import pytest
+
+from app.agents.new_chat.tools.mcp_tools_cache import (
+    CachedMCPToolDef,
+    CachedMCPTools,
+    read_cached_tools,
+)
+
+pytestmark = pytest.mark.unit
+
+
+def _make_connector(config: dict | None) -> SimpleNamespace:
+    return SimpleNamespace(id=42, config=config)
+
+
+def test_read_returns_none_when_config_is_none() -> None:
+    assert read_cached_tools(_make_connector(None)) is None
+
+
+def test_read_returns_none_when_cached_tools_missing() -> None:
+    assert read_cached_tools(_make_connector({"server_config": {}})) is None
+
+
+def test_read_returns_none_when_cached_tools_is_not_a_dict() -> None:
+    assert read_cached_tools(_make_connector({"cached_tools": []})) is None
+    assert read_cached_tools(_make_connector({"cached_tools": "stale"})) is None
+
+
+def test_read_parses_minimal_valid_payload() -> None:
+    parsed = read_cached_tools(
+        _make_connector(
+            {
+                "cached_tools": {
+                    "discovered_at": "2026-05-20T10:00:00+00:00",
+                    "tools": [
+                        {
+                            "name": "list_issues",
+                            "description": "List Linear issues",
+                            "input_schema": {"type": "object"},
+                        }
+                    ],
+                }
+            }
+        )
+    )
+    assert parsed is not None
+    assert parsed.server_version is None
+    assert parsed.server_name is None
+    assert parsed.transport is None
+    assert len(parsed.tools) == 1
+    assert parsed.tools[0].name == "list_issues"
+
+
+def test_read_parses_full_payload_with_serverinfo() -> None:
+    parsed = read_cached_tools(
+        _make_connector(
+            {
+                "cached_tools": {
+                    "discovered_at": "2026-05-20T10:00:00+00:00",
+                    "server_version": "1.2.3",
+                    "server_name": "atlassian-mcp",
+                    "transport": "streamable-http",
+                    "tools": [
+                        {"name": "create_issue", "input_schema": {}},
+                        {"name": "list_issues", "input_schema": {}},
+                    ],
+                }
+            }
+        )
+    )
+    assert parsed is not None
+    assert parsed.server_version == "1.2.3"
+    assert parsed.server_name == "atlassian-mcp"
+    assert parsed.transport == "streamable-http"
+    assert [t.name for t in parsed.tools] == ["create_issue", "list_issues"]
+
+
+def test_read_returns_none_for_corrupt_payload(caplog) -> None:
+    parsed = read_cached_tools(
+        _make_connector(
+            {
+                "cached_tools": {
+                    "discovered_at": "not-a-date",
+                    "tools": "should-be-a-list",
+                }
+            }
+        )
+    )
+    assert parsed is None
+    assert any("corrupt cached_tools" in r.getMessage() for r in caplog.records)
+
+
+def test_read_returns_none_when_tools_missing() -> None:
+    parsed = read_cached_tools(
+        _make_connector(
+            {"cached_tools": {"discovered_at": "2026-05-20T10:00:00+00:00"}}
+        )
+    )
+    assert parsed is None
+
+
+def test_tool_def_defaults_description_and_schema() -> None:
+    td = CachedMCPToolDef.model_validate({"name": "ping"})
+    assert td.description == ""
+    assert td.input_schema == {}
+
+
+def test_model_dump_json_mode_is_round_trippable() -> None:
+    original = CachedMCPTools(
+        discovered_at=datetime(2026, 5, 20, 10, 0, 0, tzinfo=UTC),
+        server_version="1.2.3",
+        server_name="atlassian-mcp",
+        transport="streamable-http",
+        tools=[CachedMCPToolDef(name="list_issues")],
+    )
+    payload = original.model_dump(mode="json")
+
+    assert payload["discovered_at"] == "2026-05-20T10:00:00Z"
+    assert payload["tools"][0]["name"] == "list_issues"
+
+    reparsed = CachedMCPTools.model_validate(payload)
+    assert reparsed.discovered_at == original.discovered_at
+    assert reparsed.tools[0].name == "list_issues"

surfsense_backend/tests/unit/agents/new_chat/tools/test_update_memory_scope.py

@@ -1,24 +1,24 @@
-"""Unit tests for memory scope validation and bullet format validation."""
+"""Unit tests for heading-based memory validation."""
 
 import pytest
 
-from app.agents.new_chat.tools.update_memory import (
-    _save_memory,
-    _validate_bullet_format,
-    _validate_memory_scope,
+from app.services.memory import MemoryScope, save_memory
+from app.services.memory.validation import (
+    validate_bullet_format,
+    validate_memory_scope,
 )
 
 pytestmark = pytest.mark.unit
 
 
-class _Recorder:
+class _FakeSession:
     def __init__(self) -> None:
-        self.applied_content: str | None = None
+        self.added = []
         self.commit_calls = 0
         self.rollback_calls = 0
 
-    def apply(self, content: str) -> None:
-        self.applied_content = content
+    def add(self, obj) -> None:
+        self.added.append(obj)
 
     async def commit(self) -> None:
         self.commit_calls += 1
@@ -27,172 +27,148 @@ class _Recorder:
         self.rollback_calls += 1
 
 
-# ---------------------------------------------------------------------------
-# _validate_memory_scope — marker-based
-# ---------------------------------------------------------------------------
-
-
-def test_validate_memory_scope_rejects_pref_marker_in_team_scope() -> None:
-    content = "- (2026-04-10) [pref] Prefers dark mode\n"
-    result = _validate_memory_scope(content, "team")
+def test_validate_memory_scope_rejects_new_personal_heading_in_team() -> None:
+    content = "## Preferences\n- 2026-04-10: Prefers dark mode\n"
+    result, _warnings = validate_memory_scope(content, "team")
     assert result is not None
     assert result["status"] == "error"
-    assert "[pref]" in result["message"]
+    assert "preferences" in result["message"]
 
 
-def test_validate_memory_scope_rejects_instr_marker_in_team_scope() -> None:
-    content = "- (2026-04-10) [instr] Always respond in Spanish\n"
-    result = _validate_memory_scope(content, "team")
-    assert result is not None
-    assert result["status"] == "error"
-    assert "[instr]" in result["message"]
+def test_validate_memory_scope_allows_old_marker_payload_in_team_scope() -> None:
+    content = "- (2026-04-10) [pref] Legacy personal marker remains readable\n"
+    result, _warnings = validate_memory_scope(content, "team")
+    assert result is None
 
 
-def test_validate_memory_scope_rejects_both_personal_markers_in_team() -> None:
+def test_validate_memory_scope_allows_team_headings() -> None:
+    content = "## Engineering Conventions\n- 2026-04-10: Uses PostgreSQL\n"
+    result, _warnings = validate_memory_scope(content, "team")
+    assert result is None
+
+
+def test_validate_bullet_format_accepts_new_and_legacy_bullets() -> None:
     content = (
-        "- (2026-04-10) [pref] Prefers dark mode\n"
-        "- (2026-04-10) [instr] Always respond in Spanish\n"
+        "## Facts\n"
+        "- 2026-04-10: Senior Python developer\n"
+        "- (2026-04-10) [fact] Legacy fact is preserved\n"
     )
-    result = _validate_memory_scope(content, "team")
-    assert result is not None
-    assert result["status"] == "error"
-    assert "[instr]" in result["message"]
-    assert "[pref]" in result["message"]
-
-
-def test_validate_memory_scope_allows_fact_in_team_scope() -> None:
-    content = "- (2026-04-10) [fact] Office is in downtown Seattle\n"
-    result = _validate_memory_scope(content, "team")
-    assert result is None
-
-
-def test_validate_memory_scope_allows_all_markers_in_user_scope() -> None:
-    content = (
-        "- (2026-04-10) [fact] Python developer\n"
-        "- (2026-04-10) [pref] Prefers concise answers\n"
-        "- (2026-04-10) [instr] Always use bullet points\n"
-    )
-    result = _validate_memory_scope(content, "user")
-    assert result is None
-
-
-def test_validate_memory_scope_allows_any_heading_in_team() -> None:
-    content = "## Architecture\n- (2026-04-10) [fact] Uses PostgreSQL for persistence\n"
-    result = _validate_memory_scope(content, "team")
-    assert result is None
-
-
-def test_validate_memory_scope_allows_any_heading_in_user() -> None:
-    content = "## My Projects\n- (2026-04-10) [fact] Working on SurfSense\n"
-    result = _validate_memory_scope(content, "user")
-    assert result is None
-
-
-# ---------------------------------------------------------------------------
-# _validate_bullet_format
-# ---------------------------------------------------------------------------
-
-
-def test_validate_bullet_format_passes_valid_bullets() -> None:
-    content = (
-        "## Work\n"
-        "- (2026-04-10) [fact] Senior Python developer\n"
-        "- (2026-04-10) [pref] Prefers dark mode\n"
-        "- (2026-04-10) [instr] Always respond in bullet points\n"
-    )
-    warnings = _validate_bullet_format(content)
+    warnings = validate_bullet_format(content)
     assert warnings == []
 
 
-def test_validate_bullet_format_warns_on_missing_marker() -> None:
-    content = "- (2026-04-10) Senior Python developer\n"
-    warnings = _validate_bullet_format(content)
+def test_validate_bullet_format_warns_on_nonstandard_bullet() -> None:
+    content = "## Facts\n- Senior Python developer\n"
+    warnings = validate_bullet_format(content)
     assert len(warnings) == 1
-    assert "Malformed bullet" in warnings[0]
-
-
-def test_validate_bullet_format_warns_on_missing_date() -> None:
-    content = "- [fact] Senior Python developer\n"
-    warnings = _validate_bullet_format(content)
-    assert len(warnings) == 1
-    assert "Malformed bullet" in warnings[0]
-
-
-def test_validate_bullet_format_warns_on_unknown_marker() -> None:
-    content = "- (2026-04-10) [context] Working on project X\n"
-    warnings = _validate_bullet_format(content)
-    assert len(warnings) == 1
-    assert "Malformed bullet" in warnings[0]
-
-
-def test_validate_bullet_format_ignores_non_bullet_lines() -> None:
-    content = "## Some Heading\nSome paragraph text\n"
-    warnings = _validate_bullet_format(content)
-    assert warnings == []
-
-
-def test_validate_bullet_format_warns_on_old_format_without_marker() -> None:
-    content = "## About the user\n- (2026-04-10) Likes cats\n"
-    warnings = _validate_bullet_format(content)
-    assert len(warnings) == 1
-
-
-# ---------------------------------------------------------------------------
-# _save_memory — end-to-end with marker scope check
-# ---------------------------------------------------------------------------
+    assert "Non-standard memory bullet" in warnings[0]
 
 
 @pytest.mark.asyncio
-async def test_save_memory_blocks_pref_in_team_before_commit() -> None:
-    recorder = _Recorder()
-    result = await _save_memory(
-        updated_memory="- (2026-04-10) [pref] Prefers dark mode\n",
-        old_memory=None,
-        llm=None,
-        apply_fn=recorder.apply,
-        commit_fn=recorder.commit,
-        rollback_fn=recorder.rollback,
-        label="team memory",
-        scope="team",
+async def test_save_memory_normalizes_legacy_marker_bullets(monkeypatch) -> None:
+    target = type("Target", (), {"memory_md": ""})()
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="- (2026-04-10) [fact] Legacy fact is preserved\n",
+        session=session,
     )
-    assert result["status"] == "error"
-    assert recorder.commit_calls == 0
-    assert recorder.applied_content is None
+
+    assert result.status == "saved"
+    assert target.memory_md == "## Memory\n- 2026-04-10: Legacy fact is preserved"
 
 
 @pytest.mark.asyncio
-async def test_save_memory_allows_fact_in_team_and_commits() -> None:
-    recorder = _Recorder()
-    content = "- (2026-04-10) [fact] Weekly standup on Mondays\n"
-    result = await _save_memory(
-        updated_memory=content,
-        old_memory=None,
-        llm=None,
-        apply_fn=recorder.apply,
-        commit_fn=recorder.commit,
-        rollback_fn=recorder.rollback,
-        label="team memory",
-        scope="team",
+async def test_save_memory_blocks_new_personal_heading_in_team_before_commit(
+    monkeypatch,
+) -> None:
+    target = type("Target", (), {"shared_memory_md": ""})()
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.TEAM,
+        target_id=1,
+        content="## Preferences\n- 2026-04-10: Prefers dark mode\n",
+        session=session,
     )
-    assert result["status"] == "saved"
-    assert recorder.commit_calls == 1
-    assert recorder.applied_content == content
+    assert result.status == "error"
+    assert session.commit_calls == 0
+    assert target.shared_memory_md == ""
 
 
 @pytest.mark.asyncio
-async def test_save_memory_includes_format_warnings() -> None:
-    recorder = _Recorder()
-    content = "- (2026-04-10) Missing marker text\n"
-    result = await _save_memory(
-        updated_memory=content,
-        old_memory=None,
-        llm=None,
-        apply_fn=recorder.apply,
-        commit_fn=recorder.commit,
-        rollback_fn=recorder.rollback,
-        label="memory",
-        scope="user",
+async def test_save_memory_allows_grandfathered_personal_heading_in_team(
+    monkeypatch,
+) -> None:
+    content = "## Preferences\n- 2026-04-10: Prefers dark mode\n"
+    target = type("Target", (), {"shared_memory_md": content})()
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.TEAM,
+        target_id=1,
+        content=content,
+        session=session,
     )
-    assert result["status"] == "saved"
-    assert "format_warnings" in result
-    assert len(result["format_warnings"]) == 1
+    assert result.status == "saved"
+    assert session.commit_calls == 1
+    assert target.shared_memory_md == content.strip()
+    assert result.warnings
+
+
+@pytest.mark.asyncio
+async def test_save_memory_strips_preamble_before_heading(monkeypatch) -> None:
+    target = type("Target", (), {"memory_md": ""})()
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="Sure, here is the update:\n\n## Facts\n- 2026-04-10: Likes cats\n",
+        session=session,
+    )
+    assert result.status == "saved"
+    assert target.memory_md == "## Facts\n- 2026-04-10: Likes cats"
+
+
+@pytest.mark.asyncio
+async def test_save_memory_rejects_long_no_heading_payload(monkeypatch) -> None:
+    target = type("Target", (), {"memory_md": ""})()
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="NO_UPDATE because there is nothing durable to remember.",
+        session=session,
+    )
+    assert result.status == "error"
+    assert "## heading" in result.message
+    assert session.commit_calls == 0

surfsense_backend/tests/unit/services/test_memory_service.py

@@ -0,0 +1,187 @@
+"""Unit tests for the first-class memory service."""
+
+from types import SimpleNamespace
+
+import pytest
+
+from app.services.memory import (
+    MemoryScope,
+    reset_memory,
+    save_memory,
+)
+
+pytestmark = pytest.mark.unit
+
+
+class _FakeSession:
+    def __init__(self) -> None:
+        self.commit_calls = 0
+        self.rollback_calls = 0
+        self.added = []
+
+    def add(self, obj) -> None:
+        self.added.append(obj)
+
+    async def commit(self) -> None:
+        self.commit_calls += 1
+
+    async def rollback(self) -> None:
+        self.rollback_calls += 1
+
+
+@pytest.mark.asyncio
+async def test_save_memory_saves_heading_based_memory(monkeypatch) -> None:
+    target = SimpleNamespace(memory_md="")
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="## Facts\n- 2026-05-19: Anish works on SurfSense\n",
+        session=session,
+    )
+
+    assert result.status == "saved"
+    assert target.memory_md.startswith("## Facts")
+    assert session.commit_calls == 1
+
+
+@pytest.mark.asyncio
+async def test_save_memory_accepts_legacy_marker_payload(monkeypatch) -> None:
+    target = SimpleNamespace(memory_md="")
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="- (2026-05-19) [fact] Legacy marker memory\n",
+        session=session,
+    )
+
+    assert result.status == "saved"
+    assert target.memory_md == "## Memory\n- 2026-05-19: Legacy marker memory"
+
+
+@pytest.mark.asyncio
+async def test_save_memory_rejects_long_no_heading_payload(monkeypatch) -> None:
+    target = SimpleNamespace(memory_md="## Facts\n- 2026-05-19: Existing\n")
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="reasoning text before NO_UPDATE should not become saved memory",
+        session=session,
+    )
+
+    assert result.status == "error"
+    assert session.commit_calls == 0
+    assert target.memory_md.startswith("## Facts")
+
+
+@pytest.mark.asyncio
+async def test_save_memory_no_update_sentinel_is_no_op(monkeypatch) -> None:
+    existing = "## Preferences\n- 2026-05-20: Existing preference\n"
+    target = SimpleNamespace(memory_md=existing)
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="NO_UPDATE",
+        session=session,
+    )
+
+    assert result.status == "no_op"
+    assert result.memory_md == existing
+    assert target.memory_md == existing
+    assert session.commit_calls == 0
+
+
+@pytest.mark.asyncio
+async def test_save_memory_no_update_sentinel_is_case_insensitive(monkeypatch) -> None:
+    existing = "## Preferences\n- 2026-05-20: Existing preference\n"
+    target = SimpleNamespace(memory_md=existing)
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        content="  no update  ",
+        session=session,
+    )
+
+    assert result.status == "no_op"
+    assert result.memory_md == existing
+    assert target.memory_md == existing
+    assert session.commit_calls == 0
+
+
+@pytest.mark.asyncio
+async def test_save_memory_grandfathers_existing_team_personal_heading(
+    monkeypatch,
+) -> None:
+    content = "## Preferences\n- 2026-05-19: Existing legacy heading\n"
+    target = SimpleNamespace(shared_memory_md=content)
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await save_memory(
+        scope=MemoryScope.TEAM,
+        target_id=1,
+        content=content,
+        session=session,
+    )
+
+    assert result.status == "saved"
+    assert result.warnings
+    assert session.commit_calls == 1
+
+
+@pytest.mark.asyncio
+async def test_reset_memory_clears_memory(monkeypatch) -> None:
+    target = SimpleNamespace(memory_md="## Facts\n- 2026-05-19: Existing\n")
+    session = _FakeSession()
+
+    async def fake_load_target(**_kwargs):
+        return target
+
+    monkeypatch.setattr("app.services.memory.service._load_target", fake_load_target)
+
+    result = await reset_memory(
+        scope=MemoryScope.USER,
+        target_id="00000000-0000-0000-0000-000000000000",
+        session=session,
+    )
+
+    assert result.status == "saved"
+    assert target.memory_md == ""

surfsense_backend/tests/unit/tasks/chat/streaming/test_stream_output.py

@@ -89,7 +89,6 @@ async def test_stream_output_emits_text_lifecycle_and_updates_result() -> None:
         "text_end:text-1",
     ]
     assert result.accumulated_text == "Hello world"
-    assert result.agent_called_update_memory is False
 
 
 async def test_stream_output_passes_runtime_context_to_agent() -> None:

surfsense_backend/tests/unit/test_stream_new_chat_contract.py

@@ -209,128 +209,6 @@ def test_stream_exception_classifies_openrouter_429_payload():
     assert extra is None
 
 
-@pytest.mark.asyncio
-async def test_preflight_swallows_non_rate_limit_errors_and_re_raises_429(monkeypatch):
-    """``_preflight_llm`` is best-effort.
-
-    - On rate-limit shaped exceptions (provider 429) it MUST re-raise so the
-      caller can drive the cooldown/repin branch.
-    - On any other transient failure it MUST swallow the error so the normal
-      stream path continues without surfacing preflight noise to the user.
-    """
-    from types import SimpleNamespace
-
-    from app.tasks.chat.stream_new_chat import _preflight_llm
-
-    class _RateLimitedError(Exception):
-        """Class-name carries 'RateLimit' so _is_provider_rate_limited triggers."""
-
-    rate_calls: list[dict] = []
-    other_calls: list[dict] = []
-
-    async def _fake_acompletion_429(**kwargs):
-        rate_calls.append(kwargs)
-        raise _RateLimitedError("simulated 429")
-
-    async def _fake_acompletion_other(**kwargs):
-        other_calls.append(kwargs)
-        raise RuntimeError("some unrelated transient failure")
-
-    fake_llm = SimpleNamespace(
-        model="openrouter/google/gemma-4-31b-it:free",
-        api_key="test",
-        api_base=None,
-    )
-
-    import litellm  # type: ignore[import-not-found]
-
-    monkeypatch.setattr(litellm, "acompletion", _fake_acompletion_429)
-    with pytest.raises(_RateLimitedError):
-        await _preflight_llm(fake_llm)
-    assert len(rate_calls) == 1
-    assert rate_calls[0]["max_tokens"] == 1
-    assert rate_calls[0]["stream"] is False
-
-    monkeypatch.setattr(litellm, "acompletion", _fake_acompletion_other)
-    # MUST NOT raise: non-rate-limit failures are swallowed.
-    await _preflight_llm(fake_llm)
-    assert len(other_calls) == 1
-
-
-@pytest.mark.asyncio
-async def test_preflight_skipped_for_auto_router_model():
-    """Router-mode ``model='auto'`` has no single deployment to ping; the
-    LiteLLM router itself owns per-deployment rate-limit accounting, so the
-    preflight helper must short-circuit instead of issuing a probe."""
-    from types import SimpleNamespace
-
-    from app.tasks.chat.stream_new_chat import _preflight_llm
-
-    fake_llm = SimpleNamespace(model="auto", api_key="x", api_base=None)
-    # Should return without raising or making any LiteLLM call.
-    await _preflight_llm(fake_llm)
-
-
-@pytest.mark.asyncio
-async def test_settle_speculative_agent_build_swallows_exceptions():
-    """``_settle_speculative_agent_build`` MUST always return cleanly so the
-    caller can safely re-touch the request-scoped session afterwards.
-
-    The helper guards the parallel preflight + agent-build path: when the
-    speculative build is being discarded (429 or non-429 preflight failure)
-    we await it solely to release any in-flight ``AsyncSession`` usage —
-    the build's outcome is irrelevant. Any exception (including
-    ``CancelledError``) leaking out would skip the caller's recovery flow
-    and re-introduce the very session-concurrency hazard the helper exists
-    to prevent.
-    """
-    import asyncio
-
-    from app.tasks.chat.stream_new_chat import _settle_speculative_agent_build
-
-    async def _raises() -> None:
-        raise RuntimeError("speculative build crashed")
-
-    async def _succeeds() -> str:
-        return "agent"
-
-    async def _slow() -> None:
-        await asyncio.sleep(0.05)
-
-    for coro in (_raises(), _succeeds(), _slow()):
-        task = asyncio.create_task(coro)
-        await _settle_speculative_agent_build(task)
-        assert task.done()
-
-
-@pytest.mark.asyncio
-async def test_settle_speculative_agent_build_handles_already_done_task():
-    """Done tasks (success or failure) must still be settled without raising."""
-    import asyncio
-
-    from app.tasks.chat.stream_new_chat import _settle_speculative_agent_build
-
-    async def _ok() -> str:
-        return "ok"
-
-    async def _bad() -> None:
-        raise ValueError("nope")
-
-    ok_task = asyncio.create_task(_ok())
-    bad_task = asyncio.create_task(_bad())
-    # Drive both to completion before settling.
-    await asyncio.sleep(0)
-    await asyncio.sleep(0)
-
-    await _settle_speculative_agent_build(ok_task)
-    await _settle_speculative_agent_build(bad_task)
-    assert ok_task.result() == "ok"
-    # ``bad_task`` exception was consumed by the settle helper; calling
-    # ``.exception()`` after the fact must still return the original error
-    # (the helper observes it but doesn't clear it).
-    assert isinstance(bad_task.exception(), ValueError)
-
-
 def test_stream_exception_classifies_thread_busy():
     exc = BusyError(request_id="thread-123")
     kind, code, severity, is_expected, user_message, extra = _classify_stream_exception(

surfsense_backend/uv.lock

@@ -7947,7 +7947,7 @@ wheels = [
 
 [[package]]
 name = "surf-new-backend"
-version = "0.0.24"
+version = "0.0.25"
 source = { editable = "." }
 dependencies = [
     { name = "alembic" },

surfsense_browser_extension/package.json

@@ -1,7 +1,7 @@
 {
 	"name": "surfsense_browser_extension",
 	"displayName": "Surfsense Browser Extension",
-	"version": "0.0.24",
+	"version": "0.0.25",
 	"description": "Extension to collect Browsing History for SurfSense.",
 	"author": "https://github.com/MODSetter",
 	"engines": {

surfsense_desktop/package.json

@@ -1,6 +1,6 @@
 {
   "name": "surfsense-desktop",
-  "version": "0.0.24",
+  "version": "0.0.25",
   "description": "SurfSense Desktop App",
   "main": "dist/main.js",
   "scripts": {

surfsense_web/app/(home)/login/GoogleLoginButton.tsx

@@ -5,6 +5,7 @@ import { Logo } from "@/components/Logo";
 import { Button } from "@/components/ui/button";
 import { trackLoginAttempt } from "@/lib/posthog/events";
 import { AmbientBackground } from "./AmbientBackground";
+import { BACKEND_URL } from "@/lib/env-config";
 
 function GoogleGLogo({ className }: { className?: string }) {
 	return (
@@ -50,7 +51,7 @@ export function GoogleLoginButton() {
 		// cross-origin fetch requests may not be sent on subsequent redirects.
 		// The authorize-redirect endpoint does a server-side redirect to Google
 		// and sets the CSRF cookie properly for same-site context.
-		window.location.href = `${process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL}/auth/google/authorize-redirect`;
+		window.location.href = `${BACKEND_URL}/auth/google/authorize-redirect`;
 	};
 	return (
 		<div className="relative w-full overflow-hidden">

surfsense_web/app/api/zero/query/route.ts

@@ -4,11 +4,9 @@ import { NextResponse } from "next/server";
 import type { Context } from "@/types/zero";
 import { queries } from "@/zero/queries";
 import { schema } from "@/zero/schema";
+import { BACKEND_URL } from "@/lib/env-config";
 
-const backendURL =
-	process.env.FASTAPI_BACKEND_INTERNAL_URL ||
-	process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL ||
-	"http://localhost:8000";
+const backendURL = BACKEND_URL;
 
 async function authenticateRequest(
 	request: Request

surfsense_web/app/dashboard/[search_space_id]/new-chat/[[...chat_id]]/page.tsx

@@ -118,7 +118,7 @@ import {
 	trackChatResponseReceived,
 } from "@/lib/posthog/events";
 import Loading from "../loading";
-
+import { BACKEND_URL } from "@/lib/env-config";
 const MobileEditorPanel = dynamic(
 	() =>
 		import("@/components/editor-panel/editor-panel").then((m) => ({
@@ -777,7 +777,7 @@ export default function NewChatPage() {
 		if (threadId) {
 			const token = getBearerToken();
 			if (token) {
-				const backendUrl = process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL || "http://localhost:8000";
+				const backendUrl = BACKEND_URL;
 				try {
 					const response = await fetch(
 						`${backendUrl}/api/v1/threads/${threadId}/cancel-active-turn`,
@@ -978,7 +978,7 @@ export default function NewChatPage() {
 			let streamBatcher: FrameBatchedUpdater | null = null;
 
 			try {
-				const backendUrl = process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL || "http://localhost:8000";
+				const backendUrl = BACKEND_URL;
 				const selection = await getAgentFilesystemSelection(searchSpaceId, {
 					localFilesystemEnabled,
 				});
@@ -1520,7 +1520,7 @@ export default function NewChatPage() {
 			}
 
 			try {
-				const backendUrl = process.env.NEXT_PUBLIC_FASTAPI_BACKEND_URL || "http://localhost:8000";
+				const backendUrl = BACKEND_URL;
 				const selection = await getAgentFilesystemSelection(searchSpaceId, {
 					localFilesystemEnabled,
 				});

surfsense_web/app/dashboard/[search_space_id]/search-space-settings/layout-shell.tsx

@@ -3,7 +3,6 @@
 import {
 	BookText,
 	Bot,
-	Brain,
 	CircleUser,
 	Earth,
 	ImageIcon,
@@ -27,7 +26,6 @@ export type SearchSpaceSettingsTab =
 	| "vision-models"
 	| "team-roles"
 	| "prompts"
-	| "team-memory"
 	| "public-links";
 
 const DEFAULT_TAB: SearchSpaceSettingsTab = "general";
@@ -89,11 +87,6 @@ export function SearchSpaceSettingsLayoutShell({
 				label: t("nav_system_instructions"),
 				icon: <BookText className="h-4 w-4" />,
 			},
-			{
-				value: "team-memory" as const,
-				label: "Team Memory",
-				icon: <Brain className="h-4 w-4" />,
-			},
 			{
 				value: "public-links" as const,
 				label: t("nav_public_links"),

surfsense_web/app/dashboard/[search_space_id]/search-space-settings/team-memory/page.tsx

@@ -1,6 +0,0 @@
-import { TeamMemoryManager } from "@/components/settings/team-memory-manager";
-
-export default async function Page({ params }: { params: Promise<{ search_space_id: string }> }) {
-	const { search_space_id } = await params;
-	return <TeamMemoryManager searchSpaceId={Number(search_space_id)} />;
-}

surfsense_web/app/dashboard/[search_space_id]/user-settings/components/MemoryContent.tsx

@@ -1,293 +0,0 @@
-"use client";
-
-import { useAtomValue } from "jotai";
-import { ArrowUp, ChevronDown, ClipboardCopy, Download, Info, Pencil } from "lucide-react";
-import { useCallback, useEffect, useRef, useState } from "react";
-import { toast } from "sonner";
-import { z } from "zod";
-import { activeSearchSpaceIdAtom } from "@/atoms/search-spaces/search-space-query.atoms";
-import { PlateEditor } from "@/components/editor/plate-editor";
-import { Alert, AlertDescription } from "@/components/ui/alert";
-import { Button } from "@/components/ui/button";
-import {
-	DropdownMenu,
-	DropdownMenuContent,
-	DropdownMenuItem,
-	DropdownMenuTrigger,
-} from "@/components/ui/dropdown-menu";
-import { Spinner } from "@/components/ui/spinner";
-
-import { baseApiService } from "@/lib/apis/base-api.service";
-
-const MEMORY_HARD_LIMIT = 25_000;
-
-const MemoryReadSchema = z.object({
-	memory_md: z.string(),
-});
-
-export function MemoryContent() {
-	const activeSearchSpaceId = useAtomValue(activeSearchSpaceIdAtom);
-	const [memory, setMemory] = useState("");
-	const [loading, setLoading] = useState(true);
-	const [saving, setSaving] = useState(false);
-	const [editQuery, setEditQuery] = useState("");
-	const [editing, setEditing] = useState(false);
-	const [showInput, setShowInput] = useState(false);
-	const textareaRef = useRef<HTMLInputElement>(null);
-	const inputContainerRef = useRef<HTMLDivElement>(null);
-
-	const fetchMemory = useCallback(async () => {
-		try {
-			setLoading(true);
-			const data = await baseApiService.get("/api/v1/users/me/memory", MemoryReadSchema);
-			setMemory(data.memory_md);
-		} catch {
-			toast.error("Failed to load memory");
-		} finally {
-			setLoading(false);
-		}
-	}, []);
-
-	useEffect(() => {
-		fetchMemory();
-	}, [fetchMemory]);
-
-	useEffect(() => {
-		if (!showInput) return;
-
-		const handlePointerDownOutside = (event: MouseEvent | TouchEvent) => {
-			const target = event.target;
-			if (!(target instanceof Node)) return;
-			if (inputContainerRef.current?.contains(target)) return;
-
-			setShowInput(false);
-		};
-
-		document.addEventListener("mousedown", handlePointerDownOutside);
-		document.addEventListener("touchstart", handlePointerDownOutside, { passive: true });
-
-		return () => {
-			document.removeEventListener("mousedown", handlePointerDownOutside);
-			document.removeEventListener("touchstart", handlePointerDownOutside);
-		};
-	}, [showInput]);
-
-	const handleClear = async () => {
-		try {
-			setSaving(true);
-			const data = await baseApiService.put("/api/v1/users/me/memory", MemoryReadSchema, {
-				body: { memory_md: "" },
-			});
-			setMemory(data.memory_md);
-			toast.success("Memory cleared");
-		} catch {
-			toast.error("Failed to clear memory");
-		} finally {
-			setSaving(false);
-		}
-	};
-
-	const handleEdit = async () => {
-		const query = editQuery.trim();
-		if (!query) return;
-
-		try {
-			setEditing(true);
-			const data = await baseApiService.post("/api/v1/users/me/memory/edit", MemoryReadSchema, {
-				body: { query, search_space_id: Number(activeSearchSpaceId) },
-			});
-			setMemory(data.memory_md);
-			setEditQuery("");
-			setShowInput(false);
-			toast.success("Memory updated");
-		} catch {
-			toast.error("Failed to edit memory");
-		} finally {
-			setEditing(false);
-		}
-	};
-
-	const openInput = () => {
-		setShowInput(true);
-		requestAnimationFrame(() => textareaRef.current?.focus());
-	};
-
-	const handleDownload = () => {
-		if (!memory) return;
-		try {
-			const blob = new Blob([memory], { type: "text/markdown;charset=utf-8" });
-			const url = URL.createObjectURL(blob);
-			const a = document.createElement("a");
-			a.href = url;
-			a.download = "personal-memory.md";
-			document.body.appendChild(a);
-			a.click();
-			document.body.removeChild(a);
-			URL.revokeObjectURL(url);
-		} catch {
-			toast.error("Failed to download memory");
-		}
-	};
-
-	const handleCopyMarkdown = async () => {
-		if (!memory) return;
-		try {
-			await navigator.clipboard.writeText(memory);
-			toast.success("Copied to clipboard");
-		} catch {
-			toast.error("Failed to copy memory");
-		}
-	};
-
-	const handleKeyDown = (e: React.KeyboardEvent<HTMLInputElement>) => {
-		if (e.key === "Enter" && !e.shiftKey) {
-			e.preventDefault();
-			handleEdit();
-		}
-	};
-
-	const displayMemory = memory.replace(/\(\d{4}-\d{2}-\d{2}\)\s*\[(fact|pref|instr)\]\s*/g, "");
-	const charCount = memory.length;
-
-	const getCounterColor = () => {
-		if (charCount > MEMORY_HARD_LIMIT) return "text-red-500";
-		if (charCount > 15_000) return "text-orange-500";
-		if (charCount > 10_000) return "text-yellow-500";
-		return "text-muted-foreground";
-	};
-
-	if (loading) {
-		return (
-			<div className="flex items-center justify-center py-12">
-				<Spinner size="md" className="text-muted-foreground" />
-			</div>
-		);
-	}
-
-	if (!memory) {
-		return (
-			<div className="flex flex-col items-center justify-center py-16 text-center">
-				<h3 className="text-base font-medium text-foreground">What does SurfSense remember?</h3>
-				<p className="mt-2 max-w-sm text-sm text-muted-foreground">
-					Nothing yet. SurfSense picks up on your preferences and context as you chat.
-				</p>
-			</div>
-		);
-	}
-
-	return (
-		<div className="space-y-4">
-			<Alert>
-				<Info />
-				<AlertDescription>
-					<p>
-						SurfSense uses this personal memory to personalize your responses across all
-						conversations.
-					</p>
-				</AlertDescription>
-			</Alert>
-
-			<div className="relative h-[380px] rounded-lg border bg-background">
-				<div className="h-full overflow-y-auto scrollbar-thin">
-					<PlateEditor
-						markdown={displayMemory}
-						readOnly
-						preset="readonly"
-						variant="default"
-						editorVariant="none"
-						className="px-5 py-4 text-sm min-h-full"
-					/>
-				</div>
-
-				{showInput ? (
-					<div className="absolute bottom-3 inset-x-3 z-10">
-						<div
-							ref={inputContainerRef}
-							className="relative flex h-[54px] items-center gap-2 rounded-[9999px] border bg-muted/60 backdrop-blur-sm pl-4 pr-1 shadow-sm"
-						>
-							<input
-								ref={textareaRef}
-								type="text"
-								value={editQuery}
-								onChange={(e) => setEditQuery(e.target.value)}
-								onKeyDown={handleKeyDown}
-								placeholder="Tell SurfSense what to remember or forget"
-								disabled={editing}
-								className="flex-1 bg-transparent text-sm outline-none placeholder:text-muted-foreground/70"
-							/>
-							<Button
-								type="button"
-								size="icon"
-								variant="ghost"
-								onClick={handleEdit}
-								disabled={editing || !editQuery.trim()}
-								className={`h-11 w-11 shrink-0 rounded-full ${
-									editing
-										? ""
-										: "bg-muted-foreground/15 hover:bg-accent hover:text-accent-foreground"
-								}`}
-							>
-								{editing ? (
-									<Spinner size="sm" />
-								) : (
-									<ArrowUp className="!h-5 !w-5 text-foreground" strokeWidth={2.25} />
-								)}
-							</Button>
-						</div>
-					</div>
-				) : (
-					<Button
-						type="button"
-						size="icon"
-						variant="secondary"
-						onClick={openInput}
-						className="absolute bottom-3 right-3 z-10 h-[54px] w-[54px] rounded-full border bg-muted/60 backdrop-blur-sm shadow-sm"
-					>
-						<Pencil className="!h-5 !w-5" />
-					</Button>
-				)}
-			</div>
-
-			<div className="flex items-center justify-between gap-2">
-				<span className={`text-xs shrink-0 ${getCounterColor()}`}>
-					{charCount.toLocaleString()} / {MEMORY_HARD_LIMIT.toLocaleString()}
-					<span className="hidden sm:inline"> characters</span>
-					<span className="sm:hidden"> chars</span>
-					{charCount > 15_000 && charCount <= MEMORY_HARD_LIMIT && " - Approaching limit"}
-					{charCount > MEMORY_HARD_LIMIT && " - Exceeds limit"}
-				</span>
-				<div className="flex items-center gap-1.5 sm:gap-2">
-					<Button
-						type="button"
-						variant="destructive"
-						size="sm"
-						className="text-xs sm:text-sm"
-						onClick={handleClear}
-						disabled={saving || editing || !memory}
-					>
-						<span className="hidden sm:inline">Reset Memory</span>
-						<span className="sm:hidden">Reset</span>
-					</Button>
-					<DropdownMenu>
-						<DropdownMenuTrigger asChild>
-							<Button type="button" variant="secondary" size="sm" disabled={!memory}>
-								Export
-								<ChevronDown className="h-3 w-3 opacity-60" />
-							</Button>
-						</DropdownMenuTrigger>
-						<DropdownMenuContent align="end">
-							<DropdownMenuItem onClick={handleCopyMarkdown}>
-								<ClipboardCopy className="h-4 w-4 mr-2" />
-								Copy as Markdown
-							</DropdownMenuItem>
-							<DropdownMenuItem onClick={handleDownload}>
-								<Download className="h-4 w-4 mr-2" />
-								Download as Markdown
-							</DropdownMenuItem>
-						</DropdownMenuContent>
-					</DropdownMenu>
-				</div>
-			</div>
-		</div>
-	);
-}

surfsense_web/app/dashboard/[search_space_id]/user-settings/layout-shell.tsx

@@ -1,7 +1,6 @@
 "use client";
 
 import {
-	Brain,
 	CircleUser,
 	Keyboard,
 	KeyRound,
@@ -26,7 +25,6 @@ export type UserSettingsTab =
 	| "api-key"
 	| "prompts"
 	| "community-prompts"
-	| "memory"
 	| "agent-permissions"
 	| "agent-status"
 	| "purchases"
@@ -75,11 +73,6 @@ export function UserSettingsLayoutShell({ searchSpaceId, children }: UserSetting
 				label: "Community Prompts",
 				icon: <Library className="h-4 w-4" />,
 			},
-			{
-				value: "memory" as const,
-				label: "Memory",
-				icon: <Brain className="h-4 w-4" />,
-			},
 			{
 				value: "agent-permissions" as const,
 				label: "Agent Permissions",