#2121st of 38 in VoiceFor voice, speech and music
realtimestt
Python speech-to-text library with VAD, wake words and realtime transcription
- Stars
- 10.2k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Aug 2026
- Language
- Python
9,482 on 31 Mar 202610,199 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
RealtimeSTT is a Python library that detects speech with WebRTC and Silero VAD and transcribes it, from a microphone or from audio chunks you feed in. The default engine is faster-whisper; sherpa-onnx, Kroko-ONNX, Moonshine, whisper.cpp, Parakeet NeMo and others are install extras. It also ships a FastAPI production server with HTTP and WebSocket endpoints and bearer-token auth.
Who it is for: Developers adding local speech-to-text or dictation to Python apps
Strengths
- Many selectable engines: faster-whisper, sherpa-onnx, Kroko-ONNX, Moonshine, whisper.cpp, Parakeet NeMo
- Documented CPU INT8 streaming profile: Nemotron for live text, Parakeet for the final transcript
- Optional wake words via Porcupine or OpenWakeWord, plus callbacks for VAD and recording events
- Packaged server binds to loopback by default; non-loopback needs bearer token and TLS
Weaknesses
- CI covers only Python 3.11 and 3.12; 3.13 and newer are not release targets
- Uses multiprocessing, so scripts need a __main__ guard, especially on Windows
- Linux needs PortAudio headers installed before the pip install
- CPU streaming recommendation is stated only for Linux x86-64
What it needs
- GPU optional
- Docker + Compose
- Needs PortAudio, Python 3.11 or 3.12, Porcupine or OpenWakeWord (optional, for wake words)
- Models: faster-whisper, sherpa-onnx, Kroko-ONNX, Moonshine, whisper.cpp
- port 8010
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 |