Laofu Voice
Speech recognition tools for applications and AI assistants, turning audio files and live audio into usable text.
Add speech recognition to your application
Laofu Voice brings audio processing, recognition routing, transcripts, and job status behind a common set of interfaces. It runs as a separate local service for integration with existing software, with MCP access for AI assistants.
Files and live audio
- Submit audio files, follow transcription progress, and export TXT, SRT, VTT, or JSON.
- Stream audio over WebSocket and receive partial and final transcripts.
- Configure local FunASR or connect Alibaba Cloud and Doubao recognition services using your own credentials.
- Integrate through HTTP, WebSocket, a Node SDK, the CLI, or MCP.
Current stage
Version 1.4.0 is an internal integration package for Linux x86_64, without a public download at present. Local recognition can work offline after the model and its dependencies are prepared. Cloud recognition requires a network connection and the relevant service account.
The current focus is speech-to-text, rather than speech synthesis or conversational voice. Recognition quality and deployment suitability need validation with audio from the intended setting.
What you can use it for
- Add audio-file transcription to an application.
- Display incremental transcripts from live audio.
- Let an AI assistant create, inspect, and export transcription jobs.