# Qwen3-TTS-12Hz-1.7B-CustomVoice

> Open-weight text-to-speech model supporting instruction-driven style control across 10 languages

Qwen3-TTS-12Hz-1.7B-CustomVoice is an open-source text-to-speech model from Alibaba's Qwen team that allows fine-grained voice style and emotional direction via natural-language prompts. It supports 10 languages and multiple built-in speaker timbres with ultra-low latency streaming synthesis. It is useful for generating realistic dialogue and character voices without robotic cadence.

Website: https://huggingface.co/Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice

## What it is for

- Synthesize expressive spoken audio directed by natural language emotion and tone prompts
- Generate multilingual character dialogue across English, Chinese, Japanese, and other languages
- Serve low-latency voice responses for real-time conversational agents and voice bots

## Features

- Supports 10 major languages (English, Chinese, Japanese, Korean, German, French, Russian, Portuguese, Spanish, Italian)
- Instruction-driven acoustic control over pitch, speed, emotion, tone, and prosody
- 9 built-in speaker timbres covering varying genders, ages, and regional dialects
- Low-latency dual-track streaming generation capable of sub-100ms first-packet audio response
- Integration with Python via the qwen-tts package and vLLM-Omni offline inference
- Bundled local Gradio web UI demo executable from the command line

## Pricing

Open source. Free and open source under Apache 2.0; commercial cloud API available via Alibaba DashScope

## Details

- Type: Web app
- Category: [AI](https://xiry.app/topics/ai.md)
- Topic: [Text to Speech](https://xiry.app/topics/ai/text-to-speech.md)
- Open source: yes
- Platforms: Linux, Windows, macOS, Code library, API, Command line
- For: Developers, AI audio researchers, game sound designers, and creators needing open-weight expressive speech synthesis
- Alternative to: ElevenLabs, F5-TTS, ChatTTS, OpenAI Audio API
- Tags: tts, speech synthesis, voice cloning, audio generation, huggingface, qwen

## Why it was saved

- [@HuggingModels, 16 February 2026](https://x.com/HuggingModels/status/2023243784637788313): Voice cloning that sounds actually real, usable in the browser and open source — plus the demo's voice-direction prompt template (gender/pitch/speed/volume/age/clarity/fluency/accent/texture/emotion/tone/personality) worth reusing.

## Compare

- [Alternatives to ElevenLabs](https://xiry.app/alternatives/elevenlabs.md)

## Related

- [ElevenLabs](https://xiry.app/resources/elevenlabs): AI voice generator, voice cloning, and conversational agent platform (Web app, Freemium)
- [Cartesia Sonic](https://xiry.app/resources/cartesia-sonic): Ultra-low-latency streaming text-to-speech and instant voice cloning model. (AI model, Freemium)
- [Dia](https://xiry.app/resources/dia): Open-source 1.6B text-to-speech model generating multi-speaker dialogue in one pass. (AI model, Open source)
- [Airy](https://xiry.app/resources/airy): Fast, low-cost text-to-speech API and studio for creators and developers (AI model, Freemium)
- [AudioPen](https://xiry.app/resources/audiopen): Converts spoken thoughts into structured, polished writing in customizable styles. (Web app, Freemium)
- [cartesia-python](https://xiry.app/resources/cartesia-python): Official Python library for Cartesia's real-time text-to-speech API (GitHub repo, Paid)

Updated 10 October 2026. Page: https://xiry.app/resources/qwen3-tts-12hz-1-7b-customvoice

Make your own library: Xiry takes your saved reels, tweets, Reddit bookmarks and Watch Later videos, makes them searchable and pulls the useful stuff out, like this library. It's invite-only for now; join the waitlist at https://xiry.app/#waitlist.