rhasspy/wyoming-faster-whisper

Wyoming protocol server for faster whisper speech to text system

What it solves

It provides a way to run the faster-whisper speech-to-text system as a network server using the Wyoming protocol, allowing other applications (like Home Assistant) to send audio and receive transcriptions locally.

How it works

The project implements a server that wraps the faster-whisper engine. It can be deployed via Docker or a local Python installation and can be configured to use various models, including those from HuggingFace, to convert speech to text.

Who it’s for

Users who want to integrate high-performance, local speech-to-text capabilities into their smart home or other applications that support the Wyoming protocol.

Highlights

  • Supports various model sizes and types, including tiny-int8 and HuggingFace models.
  • Compatible with the Home Assistant Add-on ecosystem.
  • Deployable via Docker for easy setup.
  • Uses the Wyoming protocol for network communication.