IndexTeam / Bilibili · 2024–25 · duration control

IndexTTS 2

Industrial zero-shot TTS with pinyin control for Chinese and, in IndexTTS-2, explicit duration control for dubbing. The English clip is Speaker 1 on the paper demo — same prompt as several other open systems.

Known for

Making cloned speech fit a picture cut.

Vendor / paper demo

Speaker 1 (zero-shot) · IndexTTS official demo, Speaker 1. Same sentence as XTTS, CosyVoice 2, F5-TTS, Fish Speech.

Old will is a fine fellow but poor and helpless since missus rogers had her accident.

Year / era
2024
License class
open
Cloning
Yes
Streaming
No
Emotion controls
Yes
Languages (tags)
en · zh
Official
Source
Kokoro 82M · af_heart