Skip to content

MOSS-TTSD

Bilingual dialogue speech synthesis, open source

Visit official site

What MOSS-TTSD is

The model is described as an open source bilingual dialogue speech model that turns scripts into natural speech.

What you can do with it

  • Generate Chinese and English dialogue
  • Switch speakers with zero-shot cloning
  • Produce long podcast style audio

Application scenarios

  • Make podcast audio more listenable
  • Power interactive voice answers in education
  • Give game characters realistic dialogue

Pros & cons

✓ What we like

  • Local and API inference are both offered
  • Batch tools handle several requests at once

! What to watch out for

  • The workflow needs XY Tokenizer weights
  • Input comes as JSONL files with audio references

FAQ

Which languages are named?

The material states Chinese and English dialogue speech.

Is it open source?

The material states it is fully open source and fit for commercial use.

Last reviewed: 2026-11-15

More AI audio tools tools

View all →

How we review