Fathom Ingestion Capability
Status: Read-only access confirmed
API Key: Stored securely in /home/node/.openclaw/credentials/fathom/
Created: Feb 13, 2026
What I Can Access
1. Meetings List
Endpoint: GET /external/v1/meetings
Available Data:
- ✅ Meeting title
- ✅ Recording URL
- ✅ Scheduled times (start/end)
- ✅ Actual recording times
- ✅ Calendar invitees (names, emails, domains)
- ✅ Recording ID
- ✅ Share URL
- ✅ Recorded by (name, email, team)
- ✅ Language
What's NOT in list response:
- ❌ Transcript (requires separate call)
- ❌ Summary (requires separate call)
- ❌ Action items (requires separate call)
2. Individual Meeting Details
Endpoint: GET /external/v1/meetings/{recording_id}
Additional Data:
- ✅ Full transcript (speaker + timestamp + text)
- ✅ Summary (markdown formatted)
- ✅ Action items (description + assignee + timestamp)
- ✅ CRM matches (if connected)
- ✅ Playback URLs with timestamps
3. Transcript Only
Endpoint: GET /external/v1/recordings/{recording_id}/transcript
Data:
- ✅ Speaker display names
- ✅ Text content
- ✅ Timestamps
- ✅ Matched calendar invitee email
4. Teams
Endpoint: GET /external/v1/teams
Data:
- ✅ Team list
- ✅ Team members
Current Ingestion Status
Tested: Meetings list (working ✅)
Example response:
{
"title": "ZTAG Twice-Weekly Dev Meeting",
"url": "https://fathom.video/calls/565027806",
"recording_id": 122178894,
"calendar_invitees": ["Quan Gan", "Malachi Burke", etc.],
"recorded_by": {
"name": "Quan Gan",
"email": "quan@ztag.com",
"team": "Executive"
},
"scheduled_start_time": "2026-02-13T04:00:00Z",
"recording_start_time": "2026-02-13T04:01:32Z"
}
What I Can Build (Ingestion-Only)
Phase 1: Basic Ingestion ✅ Ready
Phase 2: Automated Syncing
Phase 3: Intelligence Layer
Available Filters
Date range:
from_date (ISO 8601)
to_date (ISO 8601)
Participants:
email (filter by attendee email)
Pagination:
limit (default 100, max 500)
cursor (for next page)
Example:
curl "https://api.fathom.ai/external/v1/meetings?from_date=2026-02-01&to_date=2026-02-13&limit=100" \
-H "X-Api-Key: [KEY]"
Ingestion Strategy
Daily Batch Sync
- When: Every morning @ 6 AM PT (in morning briefing)
- What: Pull all meetings from past 24 hours
- Store:
working/meetings/YYYY-MM-DD/[recording_id].json
- Process:
- List meetings (filter: yesterday's date)
- For each meeting: fetch full details (transcript + summary + action items)
- Save as structured JSON
- Extract key insights for briefing
Real-Time Webhook (Future)
- Trigger: Fathom sends POST when recording completes
- Webhook secret: Stored in credentials
- Action: Immediately pull and process new meeting
- Notification: Alert if action items assigned to Quan/Charlie
Trust Model
Current Capability: Read-only (ingestion)
What I CAN do:
- ✅ Read all meeting data
- ✅ Pull transcripts
- ✅ Extract summaries
- ✅ Identify action items
What I CANNOT do (by design):
- ❌ Create meetings
- ❌ Edit transcripts
- ❌ Delete recordings
- ❌ Share meetings
- ❌ Modify summaries
- ❌ Mark action items complete
Elevation Path:
- Phase 1 (current): Ingestion only
- Phase 2 (future): Draft action item tracking
- Phase 3 (future, if trust earned): Auto-complete routine action items
Next Steps
Immediate (This Week):
- Build Fathom ingestion script
- Test full meeting pull (transcript + summary + actions)
- Store in structured format
- Add to morning briefing
Short-term (Next 30 Days):
- Daily batch sync operational
- Action item extraction working
- Search across meeting history
- Webhook integration for real-time
Long-term (3-6 Months):
- RAG layer includes meeting transcripts
- Proactive follow-up reminders
- Meeting prep based on past discussions
- Action item tracking automation
Created: Feb 13, 2026
Last tested: Feb 13, 2026 (API access confirmed)