#3636th of 38 in VoiceFor voice, speech and music
pyannote-audio
Python toolkit for speaker diarization with pretrained PyTorch pipelines
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 10.7k
- License
- MIT
- Last commit
- Jun 2026
- Last release
- Jun 2026
- Language
- Jupyter Notebook
9,378 on 31 Mar 202610,673 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
pyannote.audio is a PyTorch-based Python toolkit that labels who speaks when in an audio file. It loads pretrained pipelines such as community-1 from Hugging Face and runs them locally, and the pipelines and models can be fine-tuned on your own data. A separate premium precision-2 pipeline runs on pyannoteAI servers through an API key.
Who it is for: Developers adding speaker diarization to Python audio pipelines
Strengths
- Pretrained community-1 runs locally; README reports lower DER than 3.1 on 11 of 12 listed benchmarks
- Pipelines and models can be fine-tuned, with multi-GPU training via pytorch-lightning
- Python-first API with a progress hook and optional GPU placement via pipeline.to()
- MIT license, with notebook tutorials for applying, adapting and training pipelines
Weaknesses
- Needs a Hugging Face token and acceptance of user conditions to download community-1
- precision-2, with lower reported DER than community-1, is premium and needs a pyannoteAI key
- Some tutorials were written for older versions and are described as needing updates
- Optional anonymous telemetry exists; README does not say whether it is on by default
What it needs
- GPU optional
- Needs ffmpeg, torchcodec, PyTorch, Hugging Face access token, pyannoteAI API key (precision-2 only)
- Models: pyannote/speaker-diarization-community-1, pyannote/speaker-diarization-3.1, pyannote/speaker-diarization-precision-2 (hosted)
Also in Voice
See all 38| Rank | Project | Score |
|---|---|---|
| 1 | voicestudioDesktop app for voice cloning, dubbing and dictation with local speech models | 83 out of 100 |
| 2 | whisper.cppC/C++ port of OpenAI Whisper for local speech-to-text transcription | 77 out of 100 |
| 3 | VoiceboxLocal voice studio for cloning, TTS, dictation and agent speech | 71 out of 100 |
| #4 | buzzOffline desktop app for transcribing and translating audio with Whisper | 69 out of 100 |
| #5 | meetilyDesktop app that records, transcribes and summarizes meetings locally | 68 out of 100 |