Audio
Kokoro-FastAPI
Turn the beloved Kokoro voice model into an OpenAI-compatible speech API with one Docker command.
★ 5.5k starsApache-2.0
View guide →// audio · QwenLM/Qwen3-TTS
Qwen3-TTS
Qwen3-TTS is an open-source audio project by QwenLM (QwenLM/Qwen3-TTS on GitHub). It has 13,619 GitHub stars and 1,764 forks, is released under the Apache-2.0 license, and was last updated on GitHub on 17 March 2026.
Alibaba's open voice models: expressive, streaming-capable speech with free-form voice design.
★ 13.6k stars⑂ 1.8k forksApache-2.0updated 2026-03
▦ 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
- QwenLM/Qwen3-TTS
- GitHub stars
- 13,619
- Forks
- 1,764
- License
- Apache-2.0
- Last updated on GitHub
- 17 March 2026
- Category
- Audio
- Listing
- Full install guide
- Official site
- Read the docs
// how it works
How it works
- Set up a clean environment
Create an isolated Python 3.12 environment so the audio dependencies don't clash with your other projects.
- Install the package
One pip command pulls the runtime and lets you load any released Qwen3-TTS model.
- Load a model and synthesize
Pick a model size for your hardware, feed it text, and get expressive speech — with streaming for realtime use.
- Design or clone voices
Describe the voice you want in words, or provide a short sample to clone it, then reuse it across projects.
// features
Why people love it
Expressive neural voices
Natural intonation and emotion, well beyond robotic reader voices.
Streaming generation
Audio starts flowing before the full text is processed — ideal for live agents and assistants.
Free-form voice design
Describe a voice in plain language (“warm, middle-aged narrator”) and get it.
Voice cloning
Clone a voice from a short sample for consistent characters and branding.
A family of sizes
Multiple released models so you can trade quality against speed and VRAM.
Apache-2.0
Permissive licensing for personal and commercial projects.
Advertisement
// run it
Run it yourself
- Step 1 — Create the environment
Make a fresh, isolated Python 3.12 environment.
conda create -n qwen3-tts python=3.12 -y conda activate qwen3-tts - Step 2 — Install
Install the qwen-tts package from PyPI.
pip install -U qwen-tts - Step 3 — Save VRAM (optional)
Add FlashAttention 2 to cut GPU memory usage.
pip install -U flash-attn --no-build-isolation
// use cases
What you can build
- Voices for conversational AI and realtime agents.
- Multilingual narration and dubbing.
- Consistent character voices for series and games.
- Prototyping voice products without API bills.
Advertisement
// faq
Questions, answered
Qwen3-TTS or Kokoro?
Qwen3-TTS brings larger, more expressive models with streaming and voice design. Kokoro is lighter and faster to run. Choose expressive depth or lightweight speed.
Do I need a GPU?
A GPU is recommended for comfortable speeds; FlashAttention 2 substantially reduces the memory footprint.
Does it stream audio?
Yes — streaming generation is a headline feature, built for realtime applications.
Who makes it?
The Qwen team at Alibaba Cloud, the same group behind the popular Qwen language models.
Can I use it commercially?
Yes — it is released under the Apache-2.0 license.
// keep exploring
Similar tools
Local LLM
Ollama
The simplest way to run open language models on your own machine.
★ 182.1k starsMIT
View guide →Local LLM
Unsloth
Run and fine-tune LLMs and diffusion models locally — faster training, smaller memory footprint.
★ 77.2k starsApache-2.0
View guide →Audio
Chatterbox
Chatterbox — Resemble AI's open-source TTS family: expressive text-to-speech and voice cloning, multilingual V3, low-latency Turbo and tiny Nano models, MIT
★ 27.1k starsMIT
View guide →Audio
MOSS-TTS
MOSS-TTS family: open text-to-speech models for voice cloning, dialogue, voice design, and real-time agents
★ 4.2k starsApache-2.0
View guide →Audio
VoxCPM
VoxCPM2: tokenizer-free TTS for multilingual speech, creative voice design and true-to-life cloning
★ 38.4k starsApache-2.0
View guide →// alternatives
Alternatives in Audio
The most-starred other Audio projects in the LibreAI catalogue — the realistic shortlist if Qwen3-TTS 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 Qwen3-TTS — keep exploring the shelf either side of it.
- sherpa-onnx★ 15.1k
Speech-to-text, text-to-speech, speaker diarization, speech enhancement, source separation, and VAD using next-gen Kaldi with onnxruntime without Internet co…
- voice-pro★ 13.0k
Gradio WebUI for creators and developers, featuring key TTS (Edge-TTS, kokoro) and zero-shot Voice Cloning (E2 & F5-TTS, CosyVoice), with Whisper audio proce…
- F5 TTS★ 15.3k
Official code for "F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching"
- PaddleSpeech★ 12.7k
Easy-to-use Speech Toolkit including Self-Supervised Learning model, SOTA/Streaming ASR with punctuation, Streaming TTS with text…
- leon★ 17.6k
🧠 Leon is your open-source personal assistant.
- edge-tts★ 12.2k
Use Microsoft Edge's online text-to-speech service from Python WITHOUT needing Microsoft Edge or Windows or an API key
Get Qwen3-TTS
Free and open — grab the source, star the repo, and start creating.