Presets, and voices of your own
Four neural presets cover most work. When you need a specific voice, clone one from a single reference clip.
Presets and clones, addressed alike
Switching voices is changing one field in the request. Nothing else in your integration moves when you go from a preset to a voice you cloned.
- Four presets on every account. US and UK English, suited to narration, product copy and character work.
- Clones live beside them. A voice made from one reference clip is bound to a character and called exactly like a preset — see voice cloning.
- What she speaks with on camera. The voice bound to a character is the voice a live session answers in.
Your clone — one reference clip, bound to your character.
Available to every account
These are the voices the public demo serves. Rather than a mock player here, the homepage demo speaks your own words through this exact engine.
| Voice | Character | Accent | Identifier |
|---|---|---|---|
| Ava | Warm, natural | US English | en_US-lessac-high |
| Ryan | Clear, confident | US English | en_US-ryan-high |
| Alba | Bright, expressive | UK English | en_GB-alba-medium |
| Amy | Soft, conversational | US English | en_US-amy-medium |
Your own voices
A cloned voice is created from one reference clip and bound to a character. From then on it is addressed exactly like a preset — same endpoint, same parameters — so nothing in your integration changes when you switch from a preset to a clone.
If the cloning engine is unavailable, synthesis falls back to a preset rather than failing mid-conversation. Cloning works through the dashboard and the API — submit a reference and poll the voice until it reports active. Details on voice cloning.
Clone your own voice, or one you have documented, revocable permission to reproduce. There is no automated consent check — the obligation is contractual, under the acceptable use policy.
About the voices
How do I hear a preset?
Type anything into the demo on the homepage — it synthesises your own words through this exact engine, rather than playing a canned sample.
Where do cloned voices show up?
A clone is bound to a character rather than listed in a public catalogue, and it is addressed exactly like a preset — same endpoint, same parameters. How one is made is on voice cloning.
Which accents and languages are covered?
The presets cover US and UK English. A cloned voice speaks the language of its reference clip — cross-language cloning is not built.
What happens if the cloning engine is down?
Synthesis falls back to a preset voice rather than failing mid-conversation, and returns to the clone when the container is back.
Hear one now
Type anything on the homepage and it is synthesised on the spot.