IndexTTS is a zero-shot text-to-speech system that synthesizes speech by cloning a voice from a short reference audio clip. It provides independent control over speaker timbre and emotion, addressing duration control challenges in autoregressive models.
Project overview
It decouples emotional expression from speaker identity, allowing independent timbre and emotion control using an emotion reference audio file, an 8-float vector, or an emotion text script.
Project type
Audio & Speech
Use cases
Audio & Speech
Deployment
Refer to project documentation
License
License pending
Best for
Developers and AI engineers who need to integrate voice cloning and emotion control into custom scripts using a Python library or CLI.
Users who require interactive local text-to-speech synthesis with adjustable inference and acceleration settings via a WebUI.
Key capabilities
Synthesizes speech by cloning a voice using a single short reference audio clip.
Controls speech emotion independently from the speaker's voice using a separate reference audio file, an 8-float emotion vector, or an emotion text script.
Provides a local WebUI for interactive text-to-speech synthesis with adjustable inference and acceleration settings.
Offers a Python library and CLI to run text-to-speech synthesis and integrate it into custom scripts.
Limitations and risks
The pinyin control feature for pronunciation only works for valid Chinese Pinyin cases.
There are security and accuracy risks associated with unofficial websites or services claiming to be IndexTTS.
Getting started
Installation requires medium difficulty configuration: install git and git-lfs, clone the repository and pull LFS files, install the uv package manager, run uv sync --all-extras, download models to checkpoints, and run webui.py or a Python script.