granola-mcp_skill

This skill helps you retrieve Granola MCP meeting notes, quotes, and action items quickly by querying, listing, and fetching transcripts.
  • Python

2.6k

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill openclaw/skills --skill granola-mcp

  • _meta.json295 B
  • SKILL.md2.8 KB

Overview

This skill connects to Granola AI meeting notes via the MCP proxy so you can query meetings, list meetings by date range, pull full meeting details, and retrieve verbatim transcripts. It handles OAuth token refresh logic so you can recover from expired tokens and continue querying notes. Responses preserve citation links back to the original meeting notes for easy verification.

How this skill works

Use natural-language queries with query_granola_meetings to search meeting content across documents. Use list_meetings to enumerate meetings in time ranges, get_meetings to fetch full metadata and summaries for specific meeting IDs, and get_meeting_transcript to retrieve exact spoken transcripts. The skill will surface citation links (e.g., [0]) in replies so users can click through to source notes. If calls fail with auth errors, a refresh workflow updates tokens and retries.

When to use it

  • You want a quick summary of what was discussed across multiple meetings.
  • You need a chronological list of meetings in a specific date range.
  • You require full meeting details, attendees, decisions, or action items for specific meetings.
  • You need verbatim quotes or the exact transcript text from a meeting.
  • An API call fails with a 401 and you need to trigger token refresh instructions.

Best practices

  • Prefer query_granola_meetings for open-ended, natural-language questions about topics or decisions.
  • Use list_meetings to narrow results by date, then get_meetings with returned IDs for full context.
  • Use get_meeting_transcript when you need exact wording or verbatim quotes.
  • Keep OAuth credentials out of source control and follow the provided refresh script workflow when tokens expire.
  • Always retain the citation links in responses so users can access original notes for verification.

Example use cases

  • "What did we decide about the Q3 budget in the last 30 days?" — use query_granola_meetings.
  • "Show all meetings from last week and pick the ones with action items." — use list_meetings then get_meetings.
  • "Give me the verbatim transcript for meeting 123e4567-e89b-12d3-a456-426614174000." — use get_meeting_transcript.
  • "I got a 401 error from Granola — how do I refresh the token?" — run the refresh script and retry the call.
  • "Aggregate action items across meetings mentioning 'onboarding'." — use query_granola_meetings with a focused query.

FAQ

Run the provided refresh script to exchange the refresh token for a new access token, update config files, then retry the call. If refresh fails, re-authenticate via the OAuth authorize URL.

Which endpoint should I use for natural language questions?

Use query_granola_meetings for open-ended questions about topics, decisions, or action items across meetings.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational