Index TTS 2

What Index TTS 2 is good at — and what it's not

Index TTS 2 is Bilibili's voice-cloning model with a distinctive strength: fine-grained control over eight emotions. Provide a reference sample to clone a voice, then set the strength of happy, angry, sad, afraid, disgusted, melancholic, surprised, and calm — individually — to shape exactly how each line reads. Because the emotions blend, you can layer subtle combinations rather than choosing one preset feeling, which suits performed narration and dramatic dialogue.

The cloning side reproduces a specific voice from your sample, so the same character or brand voice can carry a script with a controllable emotional range. Pricing is per 1,000 characters.

Where it sits among ZOOOP's voice models: Chatterbox TTS is the voice clone built for broad multilingual coverage; LUX TTS is the cheapest clone; for preset voices use Multilingual V3. Index TTS 2's sweet spot is emotionally nuanced voice cloning.

A reasonable mental model: default to Index TTS 2 when a cloned voice needs precise emotional control, and switch to Chatterbox for many languages or LUX TTS for the lowest cost.