Fast Whisper MCP Server

🚀 High-performance Speech Recognition MCP Server Based on Faster Whisper

A high-efficiency speech recognition MCP server based on Faster Whisper, offering powerful audio transcription capabilities.

🚀 Quick Start

This is a high - efficiency speech recognition MCP server based on Faster Whisper, which provides powerful audio transcription functions.

✨ Features

• Integrates Faster Whisper for efficient speech recognition.
• Utilizes batch processing to accelerate transcription speed.
• Automatically enables CUDA acceleration if a GPU is available.
• Supports multiple model scales (from tiny to large - v3).
• Output formats include VTT subtitles, SRT, and JSON.
• Supports batch transcription of audio files in a folder.
• Caches model instances to avoid repeated loading.
• Dynamically adjusts batch size according to GPU memory.

📦 Installation

Dependencies

• Python 3.10+
• faster - whisper>=0.9.0
• torch==2.6.0+cu126
• torchaudio==2.6.0+cu126
• mcp[cli]>=1.2.0

Installation Steps

1. Clone or download this repository.
2. Create and activate a virtual environment (recommended).
3. Install dependencies:

pip install -r requirements.txt

PyTorch Installation Guide

Install the appropriate PyTorch version according to your CUDA version:

• CUDA 12.6:

pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu126

• CUDA 12.1:

pip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu121

• CPU version:

pip install torch==2.6.0 torchvision==0.21.0 torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cpu

You can check your CUDA version via nvcc --version or nvidia - smi.

💻 Usage Examples

Basic Usage

Start the Server

On Windows systems, directly run start_server.bat. On other platforms, run:

python whisper_server.py

Configure Claude Desktop

1. Open the Claude desktop configuration file:

   • Windows: %APPDATA%\Claude\claude_desktop_config.json
   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

2. Add the configuration for the Whisper server:

{
  "mcpServers": {
    "whisper": {
      "command": "python",
      "args": ["-m", "whisper_server"],
      "env": {}
    }
  }
}

3. After starting Claude desktop, the transcription function will be available.

🔧 Technical Details

Transcription Process

1. Audio Input: Supports multiple formats (e.g., wav, mp3).
2. Model Loading: Dynamically loads the Faster Whisper model.
3. Speech Recognition: Achieves efficient speech transcription.
4. Result Output: Returns a standard JSON format.

Performance Optimization Suggestions

• Use GPU acceleration to improve processing speed.
• Reasonably set the batch size to balance resource usage.
• Regularly update the model to maintain optimal performance.

📚 Documentation

Project Dependencies

• Whisper: https://github.com/openai/whisper
• Faster Whisper: https://github.com/guillaumekln/faster-whisper

Special Thanks

Thanks to the support of the following tools and teams:

• VS Code
• OpenAI's Whisper project
• GuillaumeKln's Faster Whisper implementation