Gnani Timbre v3.0 adds audio tags
Timbre v3.0 is an Indian-language text-to-speech model that uses inline tags to make a voice laugh, sigh, whisper or pause. It launches with 13 voices in 10 languages.
What it does
Timbre v3.0 is Gnani AI's expressive text-to-speech model for Indian languages. You place audio tags in angle brackets inside the script, such as <whispers>, to make the voice laugh, sigh, whisper or pause at that spot.
It launches with 13 voices, 13 audio tags and 10 languages.
Results and limits
The source is a launch guide and gives no measured results.
Worth knowing
You can try it in the Gnani AI playground. Pipecat and LiveKit support comes later, and Timbre 2.5 (without tags) stays available for everyday voice agents.
Try it yourself
<whispers> [First line of your scene]. <laughs> [Second line]. [Pause tag] [Closing line in Hindi or another supported language].