Whisper

Locally transcribes audio with whisper.cpp on Apple Silicon, supports speaker diarization, and outputs multiple formats.

typescript

5

GitHub Stars

typescript

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "smartlittleapps-local-stt-mcp": {
      "command": "node",
      "args": [
        "path/to/local-stt-mcp/mcp-server/dist/index.js"
      ],
      "env": {
        "HF_TOKEN": "YOUR_TOKEN_HERE"
      }
    }
  }
}

You run a local speech-to-text MCP server that transcribes audio on your device using whisper.cpp, optimized for Apple Silicon. It keeps all processing private, supports multiple output formats, and can diarize speakers for clear separation of voices.

How to use

Launch the local MCP server and connect your MCP client to transcribe audio files. The server handles automatic format conversion, speaker diarization when enabled, and outputs transcripts in formats like txt, json, vtt, srt, or csv. Use the available tools to transcribe short clips or long files, and verify health and version information to ensure the server is running smoothly.

How to install

Prerequisites you need to prepare before installing the MCP server.

Node.js 18+
whisper.cpp (install via brew: brew install whisper-cpp)
ffmpeg (install via brew: brew install ffmpeg) for audio format conversion
Python 3.8+ and a HuggingFace token for speaker diarization (free token)

Step-by-step setup commands you should run in your terminal.

git clone https://github.com/your-username/local-stt-mcp.git
cd local-stt-mcp/mcp-server
npm install
npm run build

# Download whisper models
npm run setup:models

# For speaker diarization, set HuggingFace token
export HF_TOKEN="your_token_here"  # Get free token from huggingface.co

Configuration and usage notes

You can configure the MCP client to connect to the local server using the following example. This config references the local MCP server started with the documented command.

{
  "mcpServers": {
    "whisper-mcp": {
      "command": "node",
      "args": ["path/to/local-stt-mcp/mcp-server/dist/index.js"]
    }
  }
}

Security and privacy notes

All processing happens locally on your device. No audio data is sent to the cloud, ensuring privacy and control over your recordings.

Troubleshooting

If the server fails to start, verify Node.js 18+ is installed and run the build step again. Check that whisper.cpp and ffmpeg are installed correctly. For speaker diarization, ensure HF_TOKEN is set and valid.

Available tools

transcribe

Basic audio transcription with automatic format conversion

transcribe_long

Long audio file processing with chunking and format conversion

transcribe_with_speakers

Speaker diarization and transcription with format support

list_models

Show available whisper models

health_check

System diagnostics

version

Server version information