GitHub radar
Qwen3-TTS: Alibaba’s browser-based text to speech
The Qwen team at Alibaba released an online demo for their Qwen3-TTS multilingual TTS system (10 languages including Russian), which allows for voice cloning with only 3 seconds of audio sample, and no login necessary!
Qwen3-TTS is the TTS model developed by Alibaba’s Qwen team. An online demo for this model is now open on Hugging Face Spaces, which supports 10 languages (Chinese, English, Japanese, Korean, German, French, Russian, Portuguese, Spanish, Italian) and their local variations. It allows users to give textual commands for specifying pitch, sentiment, and tempo to synthesize a variety of tones (e.g., happy, relaxed, exciting). It has low end-to-end latency (~97ms), meaning that you will receive almost immediate feedback. It also uses ZeroGPU, so there’s no need to register or download anything.
Why a vibe-coder should care
If you ever have to voice over a video, record a multi-lingual podcast, or read out a lengthy article, this online demo allows you to experience Alibaba’s TTS service without any hassle. With support for Russian, you can generate high-quality Russian speech based on text input; voice cloning enables you to keep using the same voice in various recordings.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Open the Qwen3-TTS demo by Alibaba: https://huggingface.co/spaces/Qwen/Qwen3-TTS — enter text in Russian and English, try different voice profiles, and show me the results.
Nothing to install — runs in the browser.
Open on Hugging Face▌ More finds