Directory / Conversation Intel / Granola

Granola

Official MCP Gate unknown Conversation Intel RESEARCHED Checked 2026-08-24

RESEARCHED Facts from public sources with URLs. No usage claims. Nobody has run this tool. The other tier is BENCH-TESTED, which means Andrew personally ran the tool on a stated date and cannot be bought at any price. Across the whole directory that count is 0.

Vendor: granola.ai · entry id 03-granola · source 03-conversation-intel.md line 296

What it does

General-purpose AI notetaker that generates enhanced meeting notes and summaries from a local desktop app.

AI features, separated from automation with an AI label on it

AI-enhanced note generation that blends your own typed notes with the transcript, plus meeting search. It is NOT a dedicated deal-scoring or CRM-writeback tool by design - it is horizontal, not sales-specific - though users can build sales workflows (e.g., drafting CRM notes from a sales call) on top of its MCP server.

RevOps role

Horizontal meeting-notes layer that a RevOps stack can wire into sales workflows via MCP, but it is not a conversation-intelligence product in the Gong/Chorus sense - no native deal scoring, no coaching scorecards.

MCP server

Status bucket

Official MCP

Auth

OAuth - no manual API key required.

Parsed URLs

3 found in the mcp_url field

The vendor ships and maintains the server itself. A wrapper built by Zapier, Composio or a similar third party does not count as official. The status was established on 2026-08-24 and has not been re-fetched since.

mcp_status, verbatim from the file:

official

mcp_url, verbatim from the file:

https://www.pulsemcp.com/servers/granola (official server at https://mcp.granola.ai/mcp); also listed at https://mcpservers.org/servers/granola-mcp

Access gate

Gate bucket

Gate unknown

Can a solo operator reach it

Not established

api_gate, verbatim from the file:

unknown/limited. No traditional public REST API or self-serve developer-pricing page was found; the MCP server is the primary documented programmatic access path. Several unofficial community MCP servers (e.g., https://github.com/chrisguillory/granola-mcp, https://github.com/Bencockin/granola-mcp) instead read the local Granola macOS app's data directly from disk - a workaround, not a sanctioned API.

83 of 293 entries carry an unknown gate. Unknown is a legal answer and it ships as unknown rather than as a guess.

API documentation

No documentation URL recorded.

263 of 293 entries are in the same position. Blank is legal and it is published as blank.

GitHub health

Not measured. github_url, github_stars, github_last_commit and github_archived are null on every entry in this build.

The refresh rail specced in SPEC section 7.2 has not been run. An unstamped star count is a lie, so nothing is shown rather than something stale.

A github.com URL already appears somewhere in this entry, which is a seed for that rail and not a measurement of repo health:

Jobs it can do

A job tag means the vendor says the tool does this. It is not a test result, not proof the capability is reachable through the tool's MCP server, and not proof it is available on the gate this entry records.

Tagged by machine-pass on 2026-08-25 against the closed 55 job vocabulary. 271 of 293 entries carry at least one tag; 827 tags are assigned in total.

Sources

4 source URLs. Raw sources field, verbatim:

https://www.pulsemcp.com/servers/granola, https://mcpservers.org/servers/granola-mcp, https://github.com/chrisguillory/granola-mcp, https://github.com/Bencockin/granola-mcp

Notes, verbatim from the file

Included per the seed list, but flagged clearly: Granola is a horizontal notetaker with an official MCP server, not a sales-specific conversation-intelligence product - do not conflate it with Gong-class tools.

Provenance

Entry id

03-granola

Source file

03-conversation-intel.md

Source line

296

Tier

RESEARCHED

last_checked

2026-08-24

Data baked

2026-08-25

Every field above is rendered from directory.json exactly as the build produced it. Nothing is summarised and nothing is dropped. The one change made at render time is typographic and it is disclosed on the methodology page.