Irodori-TTS-v4-Large

Use it to synthesize Japanese speech with a chosen voice, style, or emotion, including by conditioning generation on reference speech.

Availability

Model available on Hugging Face; inference code and installation instructions are in the GitHub repository.

Pricing model

Not established

Deployment

Inference code and installation instructions provided.

Privacy

Not established

Licence

Gemma Terms of Use and Gemma Prohibited Use Policy, plus additional ethical restrictions.

• The model supports Japanese text-to-speech, zero-shot voice cloning, text-based voice design, and style-controlled voice cloning. ^[1]

  It combines **Text (Input) + Reference Speech + Caption Text** in a unified model, enabling zero-shot voice cloning, text-based voice design, and style-controlled voice cloning.

• It accepts up to 120 seconds of combined reference audio. ^[1]

  * **Long-reference Voice Cloning:** Condition generation on one or more reference clips, with up to 120 seconds of combined reference audio.

• Emojis in the input text can control speaking styles, emotions, and non-verbal vocal expressions. ^[1]

  By inserting specific emojis into the input text, you can control speaking styles, emotions, and non-verbal vocal expressions in the generated audio.

• SilentCipher is integrated to apply invisible audio watermarks to generated outputs. ^[1]

  * **Integrated Watermarking:** Integrates [SilentCipher](https://github.com/sony/silentcipher) to apply robust, invisible audio watermarks directly to generated outputs, promoting responsible AI usage.

• The model is available on Hugging Face, with inference code and installation instructions linked from its GitHub repository.
• The model supports Japanese text input only. ^[1]

  - **Japanese Only:** This model currently supports Japanese text input only.

• ElevenLabs introduces two text-to-speech models Daily brief 2026-09-28