granola

name	granola
description	Search and read the user's public Granola meeting notes with curl commands. Use this skill to find recent meetings, read note summaries, fetch full transcripts with attendees, and browse folders — useful whenever meeting context would help with a task.

Granola: Read meeting notes via the Granola public API

Overview

Use curl commands to interact with the Granola public-notes API. This skill covers read-only access to the user's public Granola meeting notes: listing notes, reading a note's summary and transcript, and browsing folders.

Authentication assumption: The environment variable GRANOLA_API_KEY must already be set with a valid Granola public-notes API key. All commands below rely on it.

Base URL: https://public-api.granola.ai (do not use api.granola.ai — that is Granola's internal app API and rejects public-notes keys).

How to use

All requests follow this pattern:

curl -sS "<endpoint_url>" \
  -H "Authorization: Bearer $GRANOLA_API_KEY"

Required response-handling workflow

Always write API responses to a file first, then read/search only the fields you need. Transcripts in particular can be very large (a 45-minute meeting can have over a thousand segments) — never paste raw responses into chat output.

# 1) Store response
curl -sS "<endpoint_url>" \
  -H "Authorization: Bearer $GRANOLA_API_KEY" \
  -o /tmp/granola_response.json

# 2) Inspect or extract specific fields
jq '.notes[] | {id, title, created_at}' /tmp/granola_response.json
jq -r '.summary_text' /tmp/granola_response.json

Endpoints

There are three endpoints:

Method	Path	Purpose
GET	/v1/notes	List notes, newest first
GET	/v1/notes/{note_id}	Get one note (summary, attendees, optional transcript)
GET	/v1/folders	List folders

List notes

curl -sS "https://public-api.granola.ai/v1/notes?page_size=10" \
  -H "Authorization: Bearer $GRANOLA_API_KEY" \
  -o /tmp/granola_notes.json
jq '{notes: [.notes[] | {id, title, created_at, owner}], hasMore, cursor}' /tmp/granola_notes.json

Query parameters:

• page_size — integer, 1-30 (default 10)
• created_before, created_after, updated_after — date (2026-01-27) or date-time (2026-01-27T15:30:00Z)
• folder_id — folder ID (fol_...); includes notes in child folders
• cursor — opaque pagination cursor from a previous response

Each note in the response has id (not_...), title, owner ({name, email}), created_at, and updated_at. Notes are sorted newest-first by created_at.

Get a note

curl -sS "https://public-api.granola.ai/v1/notes/not_hmVtAgY9Q7hA5p" \
  -H "Authorization: Bearer $GRANOLA_API_KEY" \
  -o /tmp/granola_note.json
jq '{title, created_at, attendees, summary_text, web_url}' /tmp/granola_note.json

The response includes everything from the list entry plus:

• summary_text (always present) and summary_markdown (nullable)
• attendees — array of {name, email}
• calendar_event — nullable: {event_title, invitees, organiser, scheduled_start_time, scheduled_end_time}
• folder_membership — array of folders
• web_url — link to the note on notes.granola.ai
• transcript — null unless you pass include=transcript

Get a note with its transcript

curl -sS "https://public-api.granola.ai/v1/notes/not_hmVtAgY9Q7hA5p?include=transcript" \
  -H "Authorization: Bearer $GRANOLA_API_KEY" \
  -o /tmp/granola_transcript.json
jq '{title, segments: (.transcript | length)}' /tmp/granola_transcript.json
jq -r '.transcript[:20][] | "[\(.start_time)] \(.speaker.source): \(.text)"' /tmp/granola_transcript.json

Transcript segments have text, start_time, end_time, and speaker ({source, diarization_label} — source is e.g. microphone or system). Search the stored file with jq/rg rather than dumping all segments.

List folders

curl -sS "https://public-api.granola.ai/v1/folders?page_size=30" \
  -H "Authorization: Bearer $GRANOLA_API_KEY" \
  -o /tmp/granola_folders.json
jq '.folders[] | {id, name, parent_folder_id}' /tmp/granola_folders.json

Folders are sorted alphabetically; parent_folder_id is null for top-level folders.

Pagination

Cursor-based and identical for notes and folders: each response includes hasMore (boolean) and cursor (opaque string, null when exhausted). Pass the cursor back as the cursor query parameter to fetch the next page. Maximum page_size is 30.

Tips

• To find a meeting about a topic, list notes (paginating as needed) and match on title, then fetch the note and search summary_text before resorting to the transcript.
• Only notes the user has made public are accessible; an empty list does not mean the user has no meetings.
• A 401 response means the API key is invalid or missing; a 404 means the note ID doesn't exist or isn't public.

API Reference

Full OpenAPI specification: https://docs.granola.ai/api-reference/openapi.json

Authentication

The GRANOLA_API_KEY environment variable must be set with a Granola public-notes API key. Users can generate one in their Granola settings. Never write the API key into files, code, or chat output.