# collabora/WhisperLive

A nearly-live implementation of OpenAI's Whisper.

Repository: https://github.com/collabora/WhisperLive
Canonical: https://ross.abutalabs.com/products/whisperlive
Language: Python
License: MIT
License Family: permissive
Topics: dictation, obs, openai, text-to-speech, translation, voice-recognition, whisper, tensorrt, tensorrt-llm, whisper-tensorrt, openvino, openvino-intel, rocm
Last push: 2026-08-04T09:01:48+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 96, release rhythm 74, longevity 86
- inputs: {"age_days": 1217, "days_push": 29, "days_rel": 92, "gap_med": 76, "n_releases_24m": 8}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4241, forks 582 (observed 2026-08-28T04:08:39.614766+00:00)

## What it is
WhisperLive is a nearly-live speech-to-text application built on OpenAI's Whisper, serving real-time transcription over WebSocket and REST from microphone or file audio. It supports faster_whisper, TensorRT-LLM, and OpenVINO backends, plus features like word-level timestamps, hotwords, and speaker diarization.

## Use cases
- transcribe live microphone audio in real time
- transcribe pre-recorded audio files
- add live captions to an OBS stream
- dictate text hands-free
- translate spoken audio to another language
- run a self-hosted speech-to-text server with GPU acceleration
- diarize speakers in a meeting recording

## When to choose
- you need low-latency streaming transcription with a client-server setup
- you want GPU-optimized Whisper inference via TensorRT or OpenVINO
- you need speaker diarization or word-level timestamps alongside transcription

## When to avoid
- you only need simple offline batch transcription without a server
- you cannot install GPU toolchains or PortAudio dependencies
- you need a fully managed cloud speech API

## Facets
- artifact type: application
- maturity: active
- function: speech-recognition, http-server, websocket, llm-inference
- domain: speech-processing, artificial-intelligence
- platform: python, cross-platform
- tags: whisper, transcription, real-time, dictation, tensorrt, openvino, speaker-diarization, translation, natural-language-processing, docker, gpu, web-server

## Member repositories
- collabora/WhisperLive (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:39.614766+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:22:15.070107+00:00, confidence not recorded.
  - readme: https://github.com/collabora/WhisperLive (fetched 2026-08-28T04:08:39.614766+00:00, sha d04fa83c45e9)
- Data as of 2026-08-30T08:39:29.467469+00:00.