collabora/WhisperLive
A real-time speech-to-text transcription server built on OpenAI Whisper with support for multiple hardware backends.
Not currently ranked — collecting fresh signals.
star history
WhisperLive is a real-time transcription application that runs the OpenAI Whisper model to convert live audio or pre-recorded files into text. It provides both server and client components, supporting hardware-accelerated inference via OpenVINO and TensorRT backends. The server exposes a REST API for clients to stream audio for transcription, with features including speaker diarization, word-level timestamps, and hotword support.
Frequently asked
- What is collabora/WhisperLive?
- A real-time speech-to-text transcription server built on OpenAI Whisper with support for multiple hardware backends.
- Is WhisperLive open source?
- Yes — collabora/WhisperLive is open source, released under the MIT license.
- What language is WhisperLive written in?
- collabora/WhisperLive is primarily written in Python.
- How popular is WhisperLive?
- collabora/WhisperLive has 4.1k stars on GitHub.
- Where can I find WhisperLive?
- collabora/WhisperLive is on GitHub at https://github.com/collabora/WhisperLive.