Audio
hifi-gan
GAN vocoder that turns mel-spectrograms into high-fidelity speech audio, fast enough for real-time use.
★ 2.4k starsMIT
View guide →// audio · gradio-app/fastrtc
fastrtc
fastrtc is an open-source audio project by gradio-app (gradio-app/fastrtc on GitHub). It has 4,631 GitHub stars and 436 forks, is released under the MIT license, and was last updated on GitHub on 12 January 2026.
FastRTC turns any Python function into a real-time audio or video stream over WebRTC or WebSockets
★ 4.6k stars⑂ 436 forksMITupdated 2026-01
▦ visual — this project's README ships no embeddable screenshot,
so here's a placeholder. The guide below is complete regardless.
so here's a placeholder. The guide below is complete regardless.
// project facts
Project facts
- Repository
- gradio-app/fastrtc
- GitHub stars
- 4,631
- Forks
- 436
- License
- MIT
- Last updated on GitHub
- 12 January 2026
- Category
- Audio
- Listing
- Full install guide
- Official site
- Official site
// how it works
How it works
- Define your reply function
Write a Python generator that takes audio (or video) frames and yields responses.
- Install FastRTC
Install from PyPI, adding the vad/tts extras if you want pause detection and text-to-speech.
- Wrap it in a Stream
Create a Stream with ReplyOnPause for audio conversations, and pick a delivery mode: built-in UI, FastAPI mount, or a phone number.
- Launch and talk
Start the stream and speak to your function in real time.
// features
Why people love it
Automatic turn-taking
Built-in voice activity detection (the ReplyOnPause handler) runs your function when the user stops speaking.
WebRTC and WebSockets
Stream real-time audio or video over either transport, in send, receive, or send-receive modes.
Built-in Gradio UI
Every stream can launch with a test interface via stream.ui.launch().
FastAPI mounting
Mount a stream into an existing FastAPI application with stream.mount(app).
FastPhone
Get a temporary phone number to call your stream, using a Hugging Face token.
Extras for speech
Optional vad and tts extras add pause detection and text-to-speech support.
Advertisement
// run it
Run it yourself
- Clone the repository
Clone the project and enter its directory (library users can install from PyPI instead).
git clone https://github.com/gradio-app/fastrtc.git cd fastrtc - Install FastRTC
Install the package, with the speech extras if you need them.
pip install fastrtc pip install "fastrtc[vad, tts]" - Create and launch a stream
Wrap a reply function in a Stream and launch the built-in UI.
from fastrtc import Stream, ReplyOnPause stream = Stream(handler=ReplyOnPause(fn), modality="audio", mode="send-receive") stream.ui.launch()
// use cases
What you can build
- Build a voice assistant prototype whose backend is a single Python function
- Add real-time audio chat to an existing FastAPI product
- Test speech models interactively through the built-in UI
- Let people call an AI demo over a real phone number with FastPhone
Advertisement
// faq
Questions, answered
Who maintains FastRTC?
The Gradio team (gradio-app) — the same group behind the Gradio ML interface library.
Is FastRTC free?
Yes, it is MIT-licensed open source. FastPhone numbers are provided through Hugging Face and require an HF token.
Does it need a paid API key?
No — the library is self-contained; you supply your own speech/LLM models inside your handler function.
How does it compare to raw WebRTC libraries like aiortc?
FastRTC abstracts the transport and adds the conversational layer — VAD, turn-taking, a test UI, and FastAPI mounting — so you write a plain function instead of managing peers and tracks.
How do I get started?
`pip install fastrtc`, wrap your function with `Stream(handler=ReplyOnPause(fn), ...)`, and call `stream.ui.launch()`.
// keep exploring
Similar tools
Audio
mcp-video-analyzer
MCP server that turns any video — YouTube.
★ 85 starsMIT
View guide →Audio
gTTS
gTTS — Python library and CLI that turns text into spoken mp3 via Google Translate's speech API
★ 2.6k starsMIT
View guide →Audio
Real-Time-Voice-Cloning
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
★ 60.2k starsNOASSERTION
View guide →Audio
VieNeu-TTS
Vietnamese TTS with instant voice cloning — on-device, real-time CPU inference, 48 kHz audio
★ 2.8k starsApache-2.0
View guide →Audio
librosa
Python library for audio and music analysis
★ 8.6k starsISC
View guide →// alternatives
Alternatives in Audio
The most-starred other Audio projects in the LibreAI catalogue — the realistic shortlist if fastrtc is not the right fit.
- MoneyPrinterTurbo★ 128.3k
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
- Whisper★ 109.9k
Robust Speech Recognition via Large-Scale Weak Supervision
- GPT-SoVITS★ 62.3k
1 min voice data can also be used to train a good TTS model! (few shot voice cloning)
- Real-Time-Voice-Cloning★ 60.2k
Clone a voice from 5 seconds of audio and generate arbitrary speech in real time
- whisper.cpp★ 54.1k
Port of OpenAI's Whisper model in C/C++
// more in audio
More in Audio
Nearby Audio projects in the catalogue, closest in stars to fastrtc — keep exploring the shelf either side of it.
- Kokoro-FastAPI★ 5.5k
Turn the beloved Kokoro voice model into an OpenAI-compatible speech API with one Docker command.
- MOSS-TTS-Nano★ 4.5k
MOSS-TTS-Nano: a 0.1B open multilingual speech model that clones voices and streams in real time on CPU
- YouDub-webui★ 5.6k
Open-source AI video localization and dubbing for YouTube/Bilibili: speech recognition, subtitle translation, voice cloning, audio mixing and rendering. 开源 A…
- WhisperLive★ 4.3k
WhisperLive: a nearly-live OpenAI Whisper transcription server for files, mics, and streams
- whisper-diarization★ 5.7k
Automatic Speech Recognition with Speaker Diarization based on OpenAI Whisper
- metavoice-src★ 4.2k
MetaVoice-1B: a 1.2B-parameter foundation model for emotional, zero-shot text-to-speech
Get fastrtc
Free and open — grab the source, star the repo, and start creating.