Local inferenceEnglish + 中文

Give your words
a living voice.

Design a new voice, recreate one you have permission to use, or direct its emotion and pace—all from one bilingual model.

Shape the voice

0 / 4,000

Script What should the voice say?

Voice description Describe identity, texture, age, and delivery

Advanced controls

CFG strength 4.0 Higher values follow the voice instruction more strongly. Seed

Reuse a seed to make results easier to compare.

Your voice will land here

Audio starts playing as soon as the first generated samples arrive.

✦

A little direction goes a long way

Mention age, vocal texture, pace, mood, and speaking style. Keep the description in the same language as your script.

Running locally BreezeBlue / Breeze-TTS-2 24 kHz · research & non-commercial use