I’m building an automation (Server-to-Server OAuth app, account level) that fetches Zoom meeting content and turns it into tasks in another tool. Fetching raw transcripts via the cloud recording endpoints works fine now.
My question is specifically about AI Companion Notes, which would be a much better format for my use case than raw transcripts or meeting summaries.
What I know so far: Zoom exposes GET /meetings/meeting_summaries (account wide) and GET /meetings/{meetingId}/meeting_summary, with meeting:read:list_summaries:admin and meeting:read:summary:admin. My app also has my_notes:read:content:admin active.
What I can’t figure out:
Is there an endpoint to list and retrieve AI Companion Notes across the account, not just per meeting summaries?
Or is meeting_summaries the only path exposed today?
Has anyone successfully pulled AI Companion notes or summaries via the API in production? Any restrictions to be aware of beyond plan requirements and E2EE limitations?
Yes. Zoom now exposes an account-level route for AI Companion My Notes through GET /v2/docs/archives. Although the endpoint is filed under Canvas/Zoom Docs, its schema explicitly supports products=my_notes alongside docs and meeting_summary. Zoom added the archiving endpoints on June 8, 2026.
The important limitation is that this is an archive feed, not a live My Notes library. One note can produce multiple versions after edits or transcript changes, and results are queried by archive_time rather than meeting ID. For Server-to-Server OAuth, add docs:read:archive:admin or docs_archiving:read:admin, retain my_notes:read:content:admin, and re-authorize the app. Download URLs expire after one hour, pagination tokens after 15 minutes, page size is capped at 100, and the endpoint is not supported in the Gov cluster.
Recall.ai’s Meeting Bot API is another option for consistently capturing meeting transcripts for downstream task generation.
Thanks for confirming the /v2/docs/archives endpoint (products=my_notes).
Update: I got docs:read:archive:admin and my_notes:read:content:admin scopes added, plus the account owner granted “Notes Management” and “Notes Content” permissions in Role Management. The 403 PERMISSION_DENIED is now resolved, the endpoint returns 200.
But data and total_records stay null, even when querying a date range (April to mid-July 2026) where I can confirm at least 18 My Notes exist on the account (verified through another integration).
Is “archiving” a separate, explicit action from simply having an AI Companion note created during a meeting? Or is there an additional account-level toggle (beyond the API scope and Role Management permission) needed to actually populate the archive feed?