Local-file transcription
Use transcribe_audio to transcribe a local audio or video recording, with optional start-time stamps.
Yap MCP is a free, open-source MCP server for transcribing local recordings and searching Yap dictation history. It runs recognition on the user’s computer and provides setup instructions for compatible AI clients.
Yap MCP is an open-source MCP server that lets a compatible AI assistant transcribe local audio or video files and search the dictation history kept by the Yap app on that computer. Recognition runs locally rather than uploading recordings to a transcription service.
The official setup page supplies an npm-based command and client configuration examples. A first transcription downloads the speech model; later recognition can work offline once setup is complete. This is a developer integration package, distinct from the desktop dictation app and the browser audio-to-text page.
Use transcribe_audio to transcribe a local audio or video recording, with optional start-time stamps.
Use search_dictations to search the local history kept by Yap; the documented search is read-only.
Use prepare_speech_model to download and verify the recognition model before a transcription.
Configure the server in supported clients using the official command or client-specific instructions.
Ask a compatible assistant to transcribe an authorized local recording before summarizing it.
Search the local Yap history for a passage previously dictated on that computer.
Download the model ahead of time before relying on recognition without a network connection.
The official page lists transcribe_audio, search_dictations, and prepare_speech_model.
The setup page describes no upload, no API key, and no account for local recognition.
The page states Node.js 20 or newer, with 22.13 or newer needed to search dictations. Follow current platform-specific instructions.
The official page describes Yap MCP as free and open source. A connected AI client can have its own separate service terms and costs.