Skill
fathom
Sync your Fathom meetings once, then search, analyze, and act on them forever — offline, at scale, without burning...
When to use fathom
Choose if
You want full-history Fathom meeting analysis from an agent loop without burning API quota on every query — sync once into local SQLite, then run unlimited search, action-item rollups, account histories, and pre-call briefs offline. Alternative is the raw Fathom REST API, which forces every analytical query back through the 60 req/min rate limit.
Avoid if
You don't have a Fathom account with API access (the upstream is not free), or you need data from competing meeting-intelligence tools (Gong, Chorus, Otter, Fireflies) — coverage is Fathom-only. Also avoid for real-time live-meeting use cases; the model is sync-then- query, not push-based.
Risk Flags
- MEDIUM auth Requires FATHOM_PP_CLI_API_KEY tied to a Fathom account; missing key returns exit code 4. Fathom's pricing/tier model is not disclosed in the README — the user must already have API access.
- MEDIUM rate_limit Upstream Fathom API enforces 60 requests/minute. The CLI offers a `--throttle` flag for bulk fetching, but agents fanning out wide without it will hit the cap.
- LOW data_quality Local SQLite store can end up incomplete if the initial sync ran without `--include-transcript --include-summary --include-action-items`. Missing data requires individual re-sync by recording id.
Cost
Type: Unknown
Install
Default
npx -y @mvanhorn/printing-press install fathom
Estimated time to first success: ~10 min
Dependencies
Minimum runtime: Node.js (npx) or Go 1.26.3+ for source install; Claude Desktop 1.0.0+ for MCPB bundle
Distribution
- Repository
- https://github.com/mvanhorn/printing-press-library
- License
- MIT