Skip to content
MCP server

Whispernotes

Local MCP server and CLI for searching and transcribing voice recordings on Mac

Pricing
Freemium
Platforms
macOS, CLI, API
Version
0.2.0
Listing updated
today

Similar MCP servers

About Whispernotes

Whispernotes is a freemium MCP server for knowledge and memory. It is available on macOS, CLI, API and self-hosted.

The README says the server and CLI ship inside the Whisper Notes Mac app, so there is nothing separate to install with npm or pip. Once you enable the Local CLI & MCP toggle in the app's settings, the same binary gives you a terminal interface for listing, searching, exporting and transcribing audio or video files, and an MCP stdio server that AI clients such as Claude Desktop, Claude Code or Cursor can call.

Transcription runs on the Mac itself using the models selected in the app, such as Parakeet, SenseVoice or Whisper variants. Transcripts are stored in the local history, and the MCP transport is a Unix socket with 0600 permissions rather than a network port. Long transcripts are returned in pages so an agent can read hour-long recordings without overflowing its context.

The product suits people who record meetings, interviews or voice memos on a Mac and want to query that history from an assistant or a shell script. The CLI and MCP server are only in the direct-download DMG build, not the Mac App Store version.

Tools this server provides (6)

Taken from the project documentation; the running server may differ.

list_transcripts
Recent transcripts with ids, dates, titles
search_transcripts
Full-text matches with snippets
get_transcript
One transcript's text
get_latest_transcript
Your most recent transcript
transcribe_file
Transcribes an audio or video file, saves it to history
list_models
Installed models and which one is active

What it does

On-device transcription
Audio and video files are transcribed on the Mac, and the documentation says nothing leaves the machine.
Transcript search and reading
Full-text search across history, with options to show timestamps or speaker labels and page through long transcripts.
Batch export
Exports transcripts as txt or md files, optionally with the original audio, into a chosen folder.
MCP tools for AI clients
Exposes tools to list, search, read and transcribe recordings to any MCP-compatible client.
Local socket access control
Uses a Unix domain socket with owner-only permissions, is off by default, and clients ask approval before each tool call.
Scriptable CLI with JSON output
Prints transcripts to stdout and emits JSON with the --json flag for piping into other tools.

Good for

  • Asking Claude to summarise recent meeting transcripts and list follow-ups
  • Batch transcribing a folder of interviews or lectures from the command line
  • Searching old voice memos for a specific idea or quote
  • Exporting transcripts as Markdown notes into an Obsidian vault
  • Piping a transcript into a local LLM for summarising

Pricing

Freemium

Free for 5,000 words a week; $14 one-time payment for the full Mac app.

Prices change. Check whispernotes.app before you buy.

Whispernotes FAQ

Does the MCP server need to be installed separately?

No. The README says the MCP server is built into the Whisper Notes Mac app and the same binary provides the CLI, so there is no npx or pip package.

Can claude.ai on the web use it?

No. The listing says the claude.ai website cannot reach local MCP servers, so Claude Desktop or Claude Code should be used instead.

Which transport does the MCP server use?

The README says it uses stdio, with no HTTP endpoint, and the app's local transport is a Unix socket with 0600 permissions.

Is the CLI available in the Mac App Store version?

No. The documentation says the CLI and MCP server are only in the direct-download Mac app, not the Mac App Store build.

Is Whispernotes free?

Whispernotes has a free plan, with paid plans for more usage or features.

What are the best alternatives to Whispernotes?

The closest alternatives to Whispernotes are Macwhisper MCP Server, Whisper Memos, shyn, M1K3 and LokalBot Meetings.

Where this comes from

  • Description: written from the project's README, the project's website and its MCP Registry entry.

How AIPoweredBot checks servers