Orpheus TTS
A 3B Canopy Labs model trained on 100K hours that reaches human-level emotional expression.
Wrap your text in SSML tags for precise control:
<speak><prosody rate="slow">Slow speech</prosody></speak>
Tags the selected model understands — click to drop one into your text where it happens:
This model reads plain text, so inline tags are ignored. For tag-based emotion, switch to an expressive model like Orpheus or Bark.
Define custom pronunciations (word = pronunciation):
About Orpheus
Orpheus from Canopy Labs is a large 3-billion-parameter model built on a Llama 3.2 backbone with a SNAC audio codec, trained on over 100,000 hours of diverse speech. Its focus is emotional realism: natural emotions, emphasis, and speaking styles that can be hard to tell apart from human recordings. It is English-only but generous on length, accepting up to 2,000 characters per request, which suits audiobooks, voice acting, and other high-quality emotional narration. Because it builds on Llama 3.2, it carries the Llama 3.2 Community license and a 'Built with Llama' attribution rather than a fully permissive one.
Best for: High-quality emotional speech, audiobooks, voice acting
Browse all Orpheus voicesAt a glance
- Developer
- Canopy Labs
- License
- Llama 3.2 Community
- Tier
- standard
- Speed
- medium
- Voice cloning
- No
- Languages
- English
- Max characters
- 2000