MCP server: One MCP server.

Walkie

@adamperlisMCP server · External

Search and read your recorded meetings: notes, action items, participants, transcripts.

Transcriptionv1.6.2MIT
Freeno charge per call

One address for any MCP client. You sign in at mcp.trywalkie.com once, after adding it.

Walkie speech-to-text app on macOS showing Dictation, Read Aloud, and meeting transcripts
From the README · 1 of 3

What it does

  1. Reads
  2. Writes
  3. High impact

At most: only reads or looks up information.

What it costs

Free

Odel charges nothing per call. The publisher may charge for its own service.

What it asks for at sign-in

Access toReadChange
Meetings

From the sign-in at mcp.trywalkie.com, as the endpoint publishes it. You approve it there.

Tools

None listed · checked 10 Oct 2026

Odel could not list its tools without signing in.

Try it in the Inspector
README2 min read

Walkie app icon

Walkie for MCP

Walkie turns speech into text and meetings into searchable notes — on your computer.
This repository is Walkie's Model Context Protocol (MCP) connector, which lets Claude, Cursor, and any MCP client read your recorded meetings.

Walkie speech-to-text app on macOS showing Dictation, Read Aloud, and meeting transcripts

What Walkie does

Walkie is a speech-to-text app for macOS, Windows, and Linux:

  • Dictate anywhere. Press a global hotkey and talk — Walkie transcribes, cleans up the text (punctuation, filler removal, your custom vocabulary), and types or pastes it into whatever app you're in.
  • Record & understand meetings. Walkie records your calls and produces speaker-labeled transcripts, AI summaries, action items, and participant lists. It recognizes video calls (Meet, Zoom, Teams) and imports from Granola and Otter.
  • Local or cloud — your call. Run transcription fully on-device for privacy, or in the cloud for maximum speed, and switch right from the home screen.
  • Read Aloud. Have any text read back to you in a natural voice.
  • Yours to control. Custom dictionary, snippets, per-app text styles, and keyboard shortcuts. In Local mode, nothing leaves your machine.

Dictation is free; meetings and this MCP connector are part of the Work plan.

Walkie's dictation overlay — the Walkie grille, a live audio visualizer, a timer, and a glowing orange recording dot

Walkie recording a meeting with a live speaker-labeled transcript

What this connector adds

Once connected, your AI assistant can work with your meeting history:

  • search_meetings — find meetings by keyword across titles, your notes, and transcripts.
  • get_meeting — read one meeting's notes, action items, participants, and full transcript.
  • list_meetings — browse every meeting as metadata (requires full archive access).

All three tools are read-only — the connector never changes your data.

Install

Cursor and Grok Bot (marketplace)

The Marketplace package is walkie-plugin/ (manifest + remote MCP URL, no binaries). After listing, install from Customize → Plugins in Cursor, or Settings → Plugins → Marketplace in Grok Bot, then complete OAuth.

Until then:

  • Cursor: Customize → MCP → URL https://mcp.trywalkie.com/mcp
  • Grok Bot: tell the Bot Add this MCP server: https://mcp.trywalkie.com/mcp
  • Grok.com: connectors → New Connector → Custom → same URL

Claude Desktop (one click)

Download walkie.mcpb and open it in Claude Desktop (Settings → Extensions → Install). It launches Walkie's local server automatically and carries the Walkie icon.

Any MCP client (command)

claude mcp add walkie -- /Applications/Walkie.app/Contents/MacOS/walkie --mcp-serve

Remote (web/mobile clients)

Connect over OAuth 2.0 to https://mcp.trywalkie.com/mcp.

Requirements

  • The Walkie desktop app (macOS/Windows) — the local server reads Walkie's on-device meeting database.
  • A Walkie Work plan or higher.

Privacy

The connector runs on your computer and reads only your local Walkie data to answer your client's requests — it sends nothing anywhere itself. Full policy: https://trywalkie.com/en/privacy

License

MIT — see LICENSE.

Similar modules

Browse Transcription