SiliconFlow Voice Transcription

Provides voice transcription capabilities by processing audio files with the FunAudioLLM/SenseVoiceSmall model, returning text with confidence scores for AI workflows.

MCP-Audio is a Flask-based server that provides voice transcription capabilities by integrating with SiliconFlow's audio processing API. The server exposes endpoints for audio file uploads, processes them using the FunAudioLLM/SenseVoiceSmall model, and returns transcription results with confidence scores. It includes Docker containerization with ffmpeg support, handles base64-encoded audio input, and implements proper error handling and file management, making it suitable for applications requiring voice-to-text conversion in AI workflows.