AI Voice Cloning IndexTTS2 Online Workflow
Using the state-of-the-art IndexTTS2 large-scale voice cloning model, you can generate a highly realistic voice clone using just a sample audio clip and a text input.
How to Use?
Open your current workflow, refer to the workflow example, upload a short reference audio clip, then connect it to the audio node. Select the IndexTTS2 speech cloning model, and enter the script in the prompt input box. Click Generate and wait for the result.
FAQ?
What is the IndexTTS2 model?
IndexTTS2 is a groundbreaking autoregressive zero-shot text-to-speech system that solves the key limitation of duration control while maintaining the naturalness of speech and enhancing emotional expression. It is one of the leading speech cloning models currently available.
How Long Should the Reference Audio Be?
Generally, around 10 seconds is sufficient. Ensure the reference audio has clear vocals and no background noise for best results.
Are There Requirements for the Input Prompts?
The prompt text is the text to be cloned, so only the script needs to be entered; no other auxiliary prompts are needed. The maximum length is 500 characters.
