feat: add Codex and OpenCode routers with multi-provider support and updated configurations

demos/README.md

@@ -16,6 +16,8 @@ This directory contains demos showcasing Plano's capabilities as an AI-native pr
 | [Preference-Based Routing](llm_routing/preference_based_routing/) | Routes prompts to LLMs based on user-defined preferences and task type (e.g. code generation vs. understanding) |
 | [Model Alias Routing](llm_routing/model_alias_routing/) | Maps semantic aliases (`arch.summarize.v1`) to provider-specific models for centralized governance |
 | [Claude Code Router](llm_routing/claude_code_router/) | Extends Claude Code with multi-provider access and preference-aligned routing for coding tasks |
+| [Codex Router](llm_routing/codex_router/) | Extends Codex CLI with multi-provider access and preference-aligned routing for coding tasks |
+| [OpenCode Router](llm_routing/opencode_router/) | Extends OpenCode CLI with multi-provider access and preference-aligned routing for coding tasks |
 
 ## Agent Orchestration
 

demos/llm_routing/claude_code_router/config.yaml

@@ -19,11 +19,11 @@ model_providers:
       - name: code understanding
         description: understand and explain existing code snippets, functions, or libraries
   # Anthropic Models
-  - model: anthropic/claude-sonnet-4-5
+  - model: anthropic/claude-sonnet-4-6
-    default: true
     access_key: $ANTHROPIC_API_KEY
 
-  - model: anthropic/claude-haiku-4-5
+  - model: anthropic/claude-haiku-4-5-20251001
+    default: true
     access_key: $ANTHROPIC_API_KEY
 
   # Ollama Models
@@ -35,7 +35,7 @@ model_providers:
 model_aliases:
   # Alias for a small faster Claude model
   arch.claude.code.small.fast:
-    target: claude-haiku-4-5
+    target: claude-haiku-4-5-20251001
 
 tracing:
   random_sampling: 100

demos/llm_routing/codex_router/README.md

@@ -0,0 +1,102 @@
+# Codex Router - Multi-Model Access with Intelligent Routing
+
+Plano extends Codex to access multiple LLM providers through a single interface and route coding requests to the best configured model.
+
+## Benefits
+
+- **Single Interface**: Use Codex while routing through Plano
+- **Task-Aware Routing**: Route requests based on coding task intent
+- **Provider Flexibility**: Mix OpenAI, Anthropic, and local models behind one endpoint
+- **Routing Transparency**: Inspect exactly which model served each request
+
+## How It Works
+
+Plano sits between Codex and configured providers:
+
+```text
+Your Request -> Codex -> Plano -> Selected Model -> Response
+```
+
+## Quick Start
+
+### Prerequisites
+
+```bash
+# Install Codex CLI
+npm install -g @openai/codex
+
+# Ensure Docker is running
+docker --version
+```
+
+### 1) Enter this demo directory
+
+```bash
+cd demos/llm_routing/codex_router
+```
+
+### 2) Set API keys
+
+```bash
+export OPENAI_API_KEY="your-openai-key-here"
+export ANTHROPIC_API_KEY="your-anthropic-key-here"
+```
+
+### 3) Start Plano
+
+```bash
+# Install with uv (recommended)
+uv tool install planoai
+planoai up
+
+# Or if already installed with uv
+uvx planoai up
+```
+
+### 4) Launch Codex through Plano
+
+```bash
+planoai cli-agent codex
+# Or if installed with uv:
+uvx planoai cli-agent codex
+```
+
+The Codex launcher integration configures:
+
+```bash
+OPENAI_BASE_URL=http://127.0.0.1:12000/v1
+OPENAI_API_KEY=test
+```
+
+If `arch.codex.default` exists in `model_aliases`, `planoai cli-agent codex` automatically starts Codex with:
+
+```bash
+codex -m arch.codex.default
+```
+
+## Monitor Routing Decisions
+
+In a second terminal:
+
+```bash
+sh pretty_model_resolution.sh
+```
+
+This prints `MODEL_RESOLUTION` lines so you can see request model -> resolved model mappings in real time.
+
+## Advanced Usage
+
+### Override Codex model for a session
+
+```bash
+planoai cli-agent codex --settings='{"CODEX_MODEL":"openai/gpt-4.1-2025-04-14"}'
+```
+
+### Context window guidance
+
+Codex works best with a large context window. Use models/configuration that support at least 64k context when possible.
+
+## Notes
+
+- Plano's `default: true` model is only used when a client request does not specify a model.
+- If Codex sends an explicit model in requests, aliasing/routing rules decide the final upstream model.

demos/llm_routing/codex_router/config.yaml

@@ -0,0 +1,38 @@
+version: v0.3.0
+
+listeners:
+  - type: model
+    name: model_listener
+    port: 12000
+
+model_providers:
+  # OpenAI Models
+  - model: openai/gpt-5-2025-08-07
+    default: true
+    access_key: $OPENAI_API_KEY
+    routing_preferences:
+      - name: code generation
+        description: generating new code snippets, functions, or boilerplate based on user prompts or requirements
+
+  - model: openai/gpt-4.1-2025-04-14
+    access_key: $OPENAI_API_KEY
+    routing_preferences:
+      - name: code understanding
+        description: understand and explain existing code snippets, functions, or libraries
+
+  # Anthropic Model
+  - model: anthropic/claude-sonnet-4-6
+    access_key: $ANTHROPIC_API_KEY
+
+  # Ollama Model (optional local fallback)
+  - model: ollama/llama3.1
+    base_url: http://host.docker.internal:11434
+
+# Model aliases for Codex sessions
+model_aliases:
+  # Default model Codex should request when launched by planoai cli-agent codex
+  arch.codex.default:
+    target: gpt-5-2025-08-07
+
+tracing:
+  random_sampling: 100

demos/llm_routing/codex_router/pretty_model_resolution.sh

@@ -0,0 +1,33 @@
+#!/usr/bin/env bash
+# Pretty-print Plano MODEL_RESOLUTION lines from docker logs
+# - hides Arch-Router
+# - prints timestamp
+# - colors MODEL_RESOLUTION red
+# - colors req_model cyan
+# - colors resolved_model magenta
+# - removes provider and streaming
+
+docker logs -f plano 2>&1 \
+| awk '
+/MODEL_RESOLUTION:/ && $0 !~ /Arch-Router/ {
+  # extract timestamp between first [ and ]
+  ts=""
+  if (match($0, /\[[0-9-]+ [0-9:.]+\]/)) {
+    ts=substr($0, RSTART+1, RLENGTH-2)
+  }
+
+  # split out after MODEL_RESOLUTION:
+  n = split($0, parts, /MODEL_RESOLUTION: */)
+  line = parts[2]
+
+  # remove provider and streaming fields
+  sub(/ *provider='\''[^'\'']+'\''/, "", line)
+  sub(/ *streaming=(true|false)/, "", line)
+
+  # highlight fields
+  gsub(/req_model='\''[^'\'']+'\''/, "\033[36m&\033[0m", line)
+  gsub(/resolved_model='\''[^'\'']+'\''/, "\033[35m&\033[0m", line)
+
+  # print timestamp + MODEL_RESOLUTION
+  printf "\033[90m[%s]\033[0m \033[31mMODEL_RESOLUTION\033[0m: %s\n", ts, line
+}'

demos/llm_routing/opencode_router/README.md

@@ -0,0 +1,97 @@
+# OpenCode Router - Multi-Model Access with Intelligent Routing
+
+Plano extends OpenCode to access multiple LLM providers through a single interface and route coding requests to the best configured model.
+
+## Benefits
+
+- **Single Interface**: Use OpenCode while routing through Plano
+- **Task-Aware Routing**: Route requests based on coding task intent
+- **Provider Flexibility**: Mix OpenAI, Anthropic, and local models behind one endpoint
+- **Routing Transparency**: Inspect exactly which model served each request
+
+## How It Works
+
+Plano sits between OpenCode and configured providers:
+
+```text
+Your Request -> OpenCode -> Plano -> Selected Model -> Response
+```
+
+## Quick Start
+
+### Prerequisites
+
+- OpenCode CLI installed and available on your `PATH` (`opencode` command)
+- Docker running
+
+### 1) Enter this demo directory
+
+```bash
+cd demos/llm_routing/opencode_router
+```
+
+### 2) Set API keys
+
+```bash
+export OPENAI_API_KEY="your-openai-key-here"
+export ANTHROPIC_API_KEY="your-anthropic-key-here"
+```
+
+### 3) Start Plano
+
+```bash
+# Install with uv (recommended)
+uv tool install planoai
+planoai up
+
+# Or if already installed with uv
+uvx planoai up
+```
+
+### 4) Launch OpenCode through Plano
+
+```bash
+planoai cli-agent opencode
+# Or if installed with uv:
+uvx planoai cli-agent opencode
+```
+
+The OpenCode launcher integration configures:
+
+```bash
+OPENAI_BASE_URL=http://127.0.0.1:12000/v1
+OPENAI_API_KEY=test
+```
+
+If `arch.opencode.default` exists in `model_aliases`, `planoai cli-agent opencode` exports:
+
+```bash
+OPENAI_MODEL=<target-from-arch.opencode.default>
+```
+
+## Monitor Routing Decisions
+
+In a second terminal:
+
+```bash
+sh pretty_model_resolution.sh
+```
+
+This prints `MODEL_RESOLUTION` lines so you can see request model -> resolved model mappings in real time.
+
+## Advanced Usage
+
+### Override OpenCode model for a session
+
+```bash
+planoai cli-agent opencode --settings='{"OPENCODE_MODEL":"openai/gpt-4.1-2025-04-14"}'
+```
+
+### Context window guidance
+
+OpenCode works best with a large context window. Use models/configuration that support at least 64k context when possible.
+
+## Notes
+
+- Plano's `default: true` model is only used when a client request does not specify a model.
+- If OpenCode sends an explicit model in requests, aliasing/routing rules decide the final upstream model.

demos/llm_routing/opencode_router/config.yaml

@@ -0,0 +1,38 @@
+version: v0.3.0
+
+listeners:
+  - type: model
+    name: model_listener
+    port: 12000
+
+model_providers:
+  # OpenAI Models
+  - model: openai/gpt-5-2025-08-07
+    default: true
+    access_key: $OPENAI_API_KEY
+    routing_preferences:
+      - name: code generation
+        description: generating new code snippets, functions, or boilerplate based on user prompts or requirements
+
+  - model: openai/gpt-4.1-2025-04-14
+    access_key: $OPENAI_API_KEY
+    routing_preferences:
+      - name: code understanding
+        description: understand and explain existing code snippets, functions, or libraries
+
+  # Anthropic Model
+  - model: anthropic/claude-sonnet-4-6
+    access_key: $ANTHROPIC_API_KEY
+
+  # Ollama Model (optional local fallback)
+  - model: ollama/llama3.1
+    base_url: http://host.docker.internal:11434
+
+# Model aliases for OpenCode sessions
+model_aliases:
+  # Default model OpenCode should request when launched by planoai cli-agent opencode
+  arch.opencode.default:
+    target: gpt-5-2025-08-07
+
+tracing:
+  random_sampling: 100

demos/llm_routing/opencode_router/pretty_model_resolution.sh

@@ -0,0 +1,33 @@
+#!/usr/bin/env bash
+# Pretty-print Plano MODEL_RESOLUTION lines from docker logs
+# - hides Arch-Router
+# - prints timestamp
+# - colors MODEL_RESOLUTION red
+# - colors req_model cyan
+# - colors resolved_model magenta
+# - removes provider and streaming
+
+docker logs -f plano 2>&1 \
+| awk '
+/MODEL_RESOLUTION:/ && $0 !~ /Arch-Router/ {
+  # extract timestamp between first [ and ]
+  ts=""
+  if (match($0, /\[[0-9-]+ [0-9:.]+\]/)) {
+    ts=substr($0, RSTART+1, RLENGTH-2)
+  }
+
+  # split out after MODEL_RESOLUTION:
+  n = split($0, parts, /MODEL_RESOLUTION: */)
+  line = parts[2]
+
+  # remove provider and streaming fields
+  sub(/ *provider='\''[^'\'']+'\''/, "", line)
+  sub(/ *streaming=(true|false)/, "", line)
+
+  # highlight fields
+  gsub(/req_model='\''[^'\'']+'\''/, "\033[36m&\033[0m", line)
+  gsub(/resolved_model='\''[^'\'']+'\''/, "\033[35m&\033[0m", line)
+
+  # print timestamp + MODEL_RESOLUTION
+  printf "\033[90m[%s]\033[0m \033[31mMODEL_RESOLUTION\033[0m: %s\n", ts, line
+}'