meetings and knowledge improvements

- Limit Granola sync to 30-day lookback period
- Move Granola and Fireflies sync dirs under knowledge/Meetings/
- Note creation agent links to source meeting notes in activity entries
- Note creation agent links to Gmail threads via web URL
- Add Meetings to note type definitions with recursive tag scanning
- Tagging agent extracts meeting metadata (date, source, attendees, title, topic)
- Clicking Knowledge with no tab open auto-opens Bases view

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

apps/x/packages/core/src/knowledge/build_graph.ts

@@ -28,8 +28,8 @@ const NOTE_CREATION_AGENT = 'note_creation';
 const SYNC_INTERVAL_MS = 30 * 1000; // Check every 30 seconds
 const SOURCE_FOLDERS = [
     'gmail_sync',
-    'fireflies_transcripts',
-    'granola_notes',
+    path.join('knowledge', 'Meetings', 'fireflies'),
+    path.join('knowledge', 'Meetings', 'granola'),
 ];
 
 // Voice memos are now created directly in knowledge/Voice Memos/<date>/
@@ -192,7 +192,9 @@ async function createNotesFromBatch(
     // Add each file's content
     message += `# Source Files to Process\n\n`;
     files.forEach((file, idx) => {
-        message += `## Source File ${idx + 1}: ${path.basename(file.path)}\n\n`;
+        // Pass workspace-relative path so the agent can link back to meeting notes
+        const relativePath = path.relative(WorkDir, file.path);
+        message += `## Source File ${idx + 1}: ${relativePath}\n\n`;
         message += file.content;
         message += `\n\n---\n\n`;
     });

apps/x/packages/core/src/knowledge/granola/sync.ts

@@ -17,13 +17,14 @@ import {
 const GRANOLA_CLIENT_VERSION = '6.462.1';
 const GRANOLA_API_BASE = 'https://api.granola.ai';
 const GRANOLA_CONFIG_PATH = path.join(homedir(), 'Library', 'Application Support', 'Granola', 'supabase.json');
-const SYNC_DIR = path.join(WorkDir, 'granola_notes');
+const SYNC_DIR = path.join(WorkDir, 'knowledge', 'Meetings', 'granola');
 const STATE_FILE = path.join(SYNC_DIR, 'sync_state.json');
 const SYNC_INTERVAL_MS = 5 * 60 * 1000; // Check every 5 minutes
 const API_DELAY_MS = 1000; // 1 second delay between API calls
 const RATE_LIMIT_RETRY_DELAY_MS = 60 * 1000; // Wait 1 minute on rate limit
 const MAX_RETRIES = 3; // Maximum retries for rate-limited requests
 const MAX_BATCH_SIZE = 10; // Process max 10 documents per folder per sync
+const LOOKBACK_DAYS = 30; // Only sync documents from the last 30 days
 
 // --- Wake Signal for Immediate Sync Trigger ---
 let wakeResolve: (() => void) | null = null;
@@ -370,6 +371,10 @@ async function syncNotes(): Promise<void> {
         let hasMore = true;
         const changedTitles: string[] = [];
 
+        // Calculate lookback cutoff date
+        const lookbackCutoff = new Date();
+        lookbackCutoff.setDate(lookbackCutoff.getDate() - LOOKBACK_DAYS);
+
         // Fetch documents with pagination
         while (hasMore) {
             // Delay before API call (except first)
@@ -390,7 +395,16 @@ async function syncNotes(): Promise<void> {
             }
 
             // Process each document
+            let foundOldDoc = false;
             for (const doc of docsResponse.docs) {
+                // Skip documents outside the lookback period
+                const docDate = new Date(doc.created_at);
+                if (docDate < lookbackCutoff) {
+                    console.log(`[Granola] Document "${doc.title}" is older than ${LOOKBACK_DAYS} days, stopping pagination`);
+                    foundOldDoc = true;
+                    break;
+                }
+
                 const docUpdatedAt = doc.updated_at || doc.created_at;
                 const lastSyncedAt = state.syncedDocs[doc.id];
 
@@ -407,8 +421,15 @@ async function syncNotes(): Promise<void> {
 
                 // Convert to markdown and save
                 const markdown = documentToMarkdown(doc);
-                const filename = `${doc.id}_${cleanFilename(docTitle)}.md`;
-                const filePath = path.join(SYNC_DIR, filename);
+                const dateDir = path.join(
+                    SYNC_DIR,
+                    String(docDate.getFullYear()),
+                    String(docDate.getMonth() + 1).padStart(2, '0'),
+                    String(docDate.getDate()).padStart(2, '0')
+                );
+                ensureDir(dateDir);
+                const filename = `${cleanFilename(docTitle)}.md`;
+                const filePath = path.join(dateDir, filename);
 
                 fs.writeFileSync(filePath, markdown);
 
@@ -424,6 +445,12 @@ async function syncNotes(): Promise<void> {
                 state.syncedDocs[doc.id] = docUpdatedAt;
             }
 
+            // Stop if we hit a document outside the lookback period
+            if (foundOldDoc) {
+                hasMore = false;
+                break;
+            }
+
             // Move to next page
             offset += docsResponse.docs.length;
 

apps/x/packages/core/src/knowledge/note_creation.ts

@@ -157,6 +157,7 @@ workspace-readFile({ path: "{source_file}" })
 - Has \`Attendees:\` field
 - Has \`Meeting:\` title
 - Transcript format with speaker labels
+- Source file path is under \`knowledge/Meetings/\` (e.g. \`knowledge/Meetings/granola/...\` or \`knowledge/Meetings/fireflies/...\`)
 
 **Email indicators:**
 - Has \`From:\` and \`To:\` fields
@@ -680,6 +681,16 @@ One line summarizing this source's relevance to the entity:
 **{YYYY-MM-DD}** ({meeting|email|voice memo}): {Summary with [[links]]}
 \`\`\`
 
+**For meetings:** Include a link to the source meeting note. Derive the wiki-link path from the source file path (strip the \`.md\` extension):
+\`\`\`
+**2025-01-15** (meeting): Discussed [[Projects/Acme Integration]] timeline with [[People/David Kim]]. See [[Meetings/granola/abc123_Weekly Sync]]
+\`\`\`
+
+**For emails:** Include a Gmail web link to the thread. Extract the thread ID from the \`**Thread ID:**\` field in the email source file, then construct the URL as \`https://mail.google.com/mail/#inbox/{threadId}\`:
+\`\`\`
+**2025-01-15** (email): [[People/Sarah Chen]] sent pricing proposal for [[Projects/Acme Integration]]. [View thread](https://mail.google.com/mail/#inbox/18d5a3b2c1e4f567)
+\`\`\`
+
 **For voice memos:** Include a link to the voice memo file using the Path field:
 \`\`\`
 **2025-01-15** (voice memo): Discussed [[Projects/Acme Integration]] timeline. See [[Voice Memos/2025-01-15/voice-memo-2025-01-15T10-30-00-000Z]]
@@ -687,11 +698,13 @@ One line summarizing this source's relevance to the entity:
 
 **Important:** Use canonical names with absolute paths from resolution map in all summaries:
 \`\`\`
-# Correct (uses absolute paths):
-**2025-01-15** (meeting): [[People/Sarah Chen]] confirmed timeline with [[People/David Kim]]. Blocked on [[Topics/Security Compliance]].
+# Correct (uses absolute paths and source links):
+**2025-01-15** (meeting): [[People/Sarah Chen]] confirmed timeline with [[People/David Kim]]. Blocked on [[Topics/Security Compliance]]. See [[Meetings/fireflies/abc_Team Sync]]
+**2025-01-15** (email): [[People/Sarah Chen]] shared the contract draft. [View thread](https://mail.google.com/mail/#inbox/18d5a3b2c1e4f567)
 
-# Incorrect (uses variants or relative links):
+# Incorrect (uses variants or relative links, missing source links):
 **2025-01-15** (meeting): Sarah confirmed timeline with David. Blocked on SOC 2.
+**2025-01-15** (email): Sarah shared the contract draft.
 \`\`\`
 
 ---
@@ -888,6 +901,16 @@ ${renderNoteTypesBlock()}
 | Email (has create label) | Yes | Yes | Yes |
 | Email (only skip labels) | No (SKIP) | No | No |
 
+**Meeting activity format:** Always include a link to the source meeting note:
+\`\`\`
+**2025-01-15** (meeting): Discussed project timeline with [[People/Sarah Chen]]. See [[Meetings/granola/abc123_Weekly Sync]]
+\`\`\`
+
+**Email activity format:** Always include a Gmail web link using the Thread ID from the source:
+\`\`\`
+**2025-01-15** (email): [[People/Sarah Chen]] sent pricing proposal. [View thread](https://mail.google.com/mail/#inbox/18d5a3b2c1e4f567)
+\`\`\`
+
 **Voice memo activity format:** Always include a link to the source voice memo:
 \`\`\`
 **2025-01-15** (voice memo): Discussed project timeline with [[People/Sarah Chen]]. See [[Voice Memos/2025-01-15/voice-memo-...]]

apps/x/packages/core/src/knowledge/note_system.ts

@@ -153,6 +153,13 @@ const DEFAULT_NOTE_TYPE_DEFINITIONS: NoteTypeDefinition[] = [
     extractionGuide:
       "Look for: topic name, keywords, related people/orgs/projects, decisions, key facts",
   },
+  {
+    type: "Meetings",
+    folder: "Meetings",
+    template: "",
+    extractionGuide:
+      "Look for: meeting title, date, attendees, source (granola or fireflies), duration, topics discussed",
+  },
 ];
 
 // ── Disk-backed config with mtime caching ──────────────────────────────────

apps/x/packages/core/src/knowledge/note_tagging_agent.ts

@@ -16,14 +16,14 @@ tools:
 ---
 # Task
 
-You are a note tagging agent. Given a batch of knowledge notes (People, Organizations, Projects, Topics), you will classify each note and prepend YAML frontmatter with categorized tags and Info attributes.
+You are a note tagging agent. Given a batch of knowledge notes (People, Organizations, Projects, Topics, Meetings), you will classify each note and prepend YAML frontmatter with categorized tags and Info/metadata attributes.
 
 # Instructions
 
 1. For each note file provided in the message, read its content carefully.
-2. Determine the note type from its folder path (People/, Organizations/, Projects/, Topics/).
+2. Determine the note type from its folder path (People/, Organizations/, Projects/, Topics/, Meetings/).
 3. Classify the note using the Rowboat Tag System (Note Tags section) appended below.
-4. Extract attributes from the note's \`## Info\` section (or \`## About\` for Topics).
+4. Extract attributes from the note's \`## Info\` section (or \`## About\` for Topics). For Meetings, extract metadata from the note content and file path (see Meeting extraction rules below).
 5. Use \`workspace-edit\` to prepend YAML frontmatter to the file. The oldString should be the first line of the file (the \`# Title\` heading), and the newString should be the frontmatter followed by that same first line.
 6. If the note already has frontmatter (starts with \`---\`), skip it.
 
@@ -97,6 +97,12 @@ Extract all \`**Key:** value\` fields from the \`## Info\` (or \`## About\`) sec
 - **Organizations**: type, industry, relationship, domain, aliases, first_met, last_seen
 - **Projects**: type, status, started, last_activity
 - **Topics** (from \`## About\`): keywords, aliases, first_mentioned, last_mentioned
+- **Meetings**: Extract from the note content and file path:
+  - \`date\`: meeting date (from the file path \`Meetings/{source}/YYYY/MM/DD/\` or from \`created_at\`/\`Date:\` in content)
+  - \`source\`: \`granola\` or \`fireflies\` (from the file path)
+  - \`attendees\`: list of attendee names (from \`Attendees:\` field or participant list)
+  - \`title\`: meeting title
+  - \`topic\`: relevant topic tags based on meeting content
 
 Note: For Organizations, the Info \`**Relationship:**\` field is separate from the \`relationship\` tag category. Include both — the Info field as \`info_relationship\` and the tag as \`relationship\`.
 
@@ -122,7 +128,11 @@ Note: For Organizations, the Info \`**Relationship:**\` field is separate from t
 7. **For Topic notes**, include:
    - The relevant topic tag
    - Source tags
-8. **Only use tags from the Rowboat Tag System** — do not invent new tags.
+8. **For Meeting notes**, include:
+   - \`source: meeting\`
+   - Topic tags based on what was discussed
+   - The \`date\`, \`attendees\`, and \`title\` fields extracted from content
+9. **Only use tags from the Rowboat Tag System** — do not invent new tags.
 9. Process all files in the batch. Do not skip any unless they already have frontmatter.
 
 ---

apps/x/packages/core/src/knowledge/sync_fireflies.ts

@@ -6,7 +6,7 @@ import { serviceLogger, type ServiceRunContext } from '../services/service_logge
 import { limitEventItems } from './limit_event_items.js';
 
 // Configuration
-const SYNC_DIR = path.join(WorkDir, 'fireflies_transcripts');
+const SYNC_DIR = path.join(WorkDir, 'knowledge', 'Meetings', 'fireflies');
 const SYNC_INTERVAL_MS = 30 * 60 * 1000; // Check every 30 minutes (reduced from 1 minute)
 const STATE_FILE = path.join(SYNC_DIR, 'sync_state.json');
 const LOOKBACK_DAYS = 30; // Last 1 month
@@ -569,8 +569,16 @@ async function syncMeetings() {
                 
                 // Convert to markdown and save
                 const markdown = meetingToMarkdown(meetingData);
-                const filename = `${meetingId}_${cleanFilename(meetingData.title || 'untitled')}.md`;
-                const filePath = path.join(SYNC_DIR, filename);
+                const meetingDate = new Date(meetingData.dateString || meetingData.date || Date.now());
+                const dateDir = path.join(
+                    SYNC_DIR,
+                    String(meetingDate.getFullYear()),
+                    String(meetingDate.getMonth() + 1).padStart(2, '0'),
+                    String(meetingDate.getDate()).padStart(2, '0')
+                );
+                fs.mkdirSync(dateDir, { recursive: true });
+                const filename = `${cleanFilename(meetingData.title || 'untitled')}.md`;
+                const filePath = path.join(dateDir, filename);
                 
                 fs.writeFileSync(filePath, markdown);
                 console.log(`[Fireflies] Saved: ${filename}`);

apps/x/packages/core/src/knowledge/tag_notes.ts

@@ -30,17 +30,17 @@ function getUntaggedNotes(state: NoteTaggingState): string[] {
     const untagged: string[] = [];
     const noteFolders = getNoteTypeDefinitions().map(d => d.folder);
 
-    for (const folder of noteFolders) {
-        const folderPath = path.join(KNOWLEDGE_DIR, folder);
-        if (!fs.existsSync(folderPath)) {
-            continue;
-        }
-
-        const entries = fs.readdirSync(folderPath);
+    function scanDir(dir: string) {
+        const entries = fs.readdirSync(dir);
         for (const entry of entries) {
-            const fullPath = path.join(folderPath, entry);
+            const fullPath = path.join(dir, entry);
             const stat = fs.statSync(fullPath);
 
+            if (stat.isDirectory()) {
+                scanDir(fullPath);
+                continue;
+            }
+
             if (!stat.isFile() || !entry.endsWith('.md')) {
                 continue;
             }
@@ -64,6 +64,14 @@ function getUntaggedNotes(state: NoteTaggingState): string[] {
         }
     }
 
+    for (const folder of noteFolders) {
+        const folderPath = path.join(KNOWLEDGE_DIR, folder);
+        if (!fs.existsSync(folderPath)) {
+            continue;
+        }
+        scanDir(folderPath);
+    }
+
     return untagged;
 }