Orpheus - SLNG Documentation

Orpheus by CanopyLabs is a low-latency text-to-speech model hosted on SLNG infrastructure. It supports 8 English voices and built-in emotion control, letting you shape how the voice sounds without post-processing. Select a voice by passing its voice ID in your request body. Optionally set an emotion to change the delivery style.

{
  "voice": "tara",
  "prompt": "Hello from SLNG and Orpheus.",
  "emotion": "happy"
}

Emotion control

Orpheus supports an emotion parameter that changes the delivery style of any voice. When omitted, the voice uses its default neutral tone.

Emotion	Description	Sample
`neutral`	Default, balanced delivery
`happy`	Upbeat, warm, and positive
`sad`	Subdued, empathetic tone
`angry`	Firm, intense delivery
`surprised`	Excited, astonished tone

Voices

Voice ID	Name	Sample	Gender	Age	Tone	Use Case
`dan`	Dan		masculine	Adult	Friendly, Casual	Tutorials, Explainer Videos, Chatbots
`jess`	Jess		feminine	Young Adult	Energetic, Youthful	Social Content, Ads, Upbeat Narration
`leah`	Leah		feminine	Adult	Warm, Gentle	Wellness Apps, Audiobooks, Meditation
`leo`	Leo		masculine	Adult	Authoritative, Deep	News, Corporate Narration, Documentaries
`mia`	Mia		feminine	Adult	Professional, Articulate	E-learning, Corporate Training, Presentations
`tara`	Tara		feminine	Adult	Conversational, Clear	Podcasts, AI Assistants, Interviews
`zac`	Zac		masculine	Adult	Enthusiastic, Dynamic	Sports, Gaming, Promos
`zoe`	Zoe		feminine	Adult	Calm, Soothing	Relaxation Apps, Gentle Narration

⌘I