Skip to content
MCP server

Yap

Private on-device speech to text for AI assistants: transcribe files, search dictations

Runs locallyStatus and history
Install
Pricing
Open source
Platforms
macOS, Windows, Linux
Runs
Locally
Version
1.0.0
Listing updated
2 days ago

About Yap

Yap is an open-source MCP server for design and media. It runs locally on your machine and is started by your AI client.

Yap MCP runs speech recognition on your own computer, using NVIDIA Parakeet TDT v3 through sherpa-onnx on the CPU. Audio is read from the path you give, decoded in memory and discarded, so nothing is uploaded, logged or stored. The only network request is a one-time, checksum-verified download of the speech model, about 670 MB. After that it works offline.

It handles WAV, MP3, M4A/AAC, MP4, FLAC, OGG, Opus, WebM, AIFF, CAF and more, of any length, and detects the language among 25 European languages. The author measured about ten times faster than real time on a recent laptop. Optional start-time stamps and Yap's cleanup (capitalisation, punctuation, filler words removed) are available, and a yap-transcribe command does the same job in a terminal.

The search tool reads the dictation history that the separate Yap dictation app keeps locally, so it only finds results once you have dictated with that app. The MCP server is MIT-licensed and needs Node.js 20 or newer (22.13 or newer for dictation search). The Yap app itself is a paid product with a short free period.

Install

Run in a terminal

claude mcp add yap -- npx -y https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb

Add to claude_desktop_config.json

{
  "mcpServers": {
    "yap": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "yap": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "yap": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "yap": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "yap": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb"
      ]
    }
  }
}

Tools this server provides (3)

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

transcribe_audio
Transcribes a local audio or video file, with optional timestamps and cleanup.
search_dictations
Searches the local Yap dictation history, newest first, read-only.
prepare_speech_model
Downloads and verifies the speech model or reports its status.

What it does

Local transcription
Transcribes audio and video files on your CPU with no upload, key or account.
Timestamps and cleanup
Optional start-time stamps per passage and text cleanup.
25 European languages
Language is detected automatically among 25 languages.
Dictation search
Searches the Yap app's local history, newest first, read-only.
Model preparation
Downloads and verifies the speech model before first use, or reports its status.
Many install routes
One-click bundle for Claude Desktop, a Claude Code plugin, a Gemini CLI extension and a plain npx command.

Good for

  • Transcribe an interview recording with timestamps
  • Summarise a voice memo sitting on your desktop
  • Find what you dictated about a topic this week
  • Transcribe sensitive audio without sending it to a cloud service
  • Batch-transcribe several files from the terminal

Pricing

Open source

MCP server is free (MIT); the separate Yap dictation app is $6.90/month or $69/year

Prices change. Check yapapp.cc before you buy.

Yap FAQ

Does my audio leave my computer?

No. It is decoded in memory and discarded; the text then goes to the assistant you asked, like anything else you show it.

Do I need the Yap app?

Not for transcription. Only search_dictations needs the app's history.

Which languages work?

25 European languages including English, German, French, Spanish, Greek and Maltese.

What does it need?

Node.js 20 or newer, on macOS, Windows x64 or Linux x64 and arm64.

What licence is it?

The server is MIT-licensed and the speech model is CC-BY-4.0.

Is Yap free?

Yes. Yap is open source, so you can use and self-host it for free.

How do I add Yap to Claude Code?

Run this command in your terminal: claude mcp add yap -- npx -y https://github.com/founden-bot/yap-mcp/releases/download/v1.0.0/yap.mcpb

What are the best alternatives to Yap?

The closest alternatives to Yap are HyperFrames by HeyGen, Ha MCP, ffmpeg-skill, Concord MCP and note1.

Alternatives to Yap

  1. Hosted MCP connector for creating and rendering videos from HTML through AI chat

    • Remote
    • Freemium
    • 85.3K GitHub stars
    • Design and media
  2. Ha MCP

    MCP server

    MCP server for managing and automating Home Assistant from AI assistants

    Runs locally
    • 8 tools
    • Open source
    • 7.9K GitHub stars
    • Productivity
  3. ffmpeg-skill

    MCP server

    Local FFmpeg video and audio editing tools for coding agents over MCP

    Runs locally
    • Open source
    • 6.9K GitHub stars
    • Design and media
  4. Concord MCP

    MCP server

    Local-first MCP server for messaging and shared work-state between coding agents

    Runs locally
    • Freemium
    • 1.3K GitHub stars
    • Developer tools
  5. note1

    MCP server

    MCP server giving AI assistants search and summaries across your note1 meetings

    • 4 tools
    • Remote
    • Free trial
    • 1 GitHub star
    • Productivity