Voices

List available voices, and clone or design a new one from an audio sample or a prompt.

Voices power audio generation. You can use your own cloned voices, your team’s, or system presets — and create new ones by cloning from a sample or designing from a prompt.

List voices

GET /v1/voices

Returns your and your team’s voices plus system presets. The cursor paginates only the owned/team collection; the full preset catalog repeats on every page so you never merge partial catalogs.

Query param	Default	Description
`presets`	`true`	Set `false` to exclude system presets
`cursor`	—	Opaque cursor from the previous page — do not construct it
`limit`	all (legacy)	Owned/team voices per page (1–100); supplying it activates pagination

{
  "data": {
    "voices": [
      { "id": "uuid", "name": "My Voice", "status": "ready", "source_type": "cloned" }
    ],
    "presets": [
      { "id": "uuid", "name": "Aria", "status": "ready", "source_type": "preset" }
    ]
  },
  "meta": { "has_more": false }
}

Get a voice

GET /v1/voices/:id

Accessible for your own voices and presets. A voice owned by another user returns 403.

Clone or design a voice

POST /v1/voices/clone

Asynchronous — returns 201 with status: "processing". Poll GET /v1/voices/:id until ready or failed. Requires an Idempotency-Key.

Field	Type	Required	Default	Description
`name`	string	yes	—	1–100 characters
`description`	string	no	—	Max 500 characters
`source_url`	string (URL)	yes for clone	—	An owned Someone HTTPS audio sample
`model`	enum	no	`minimax`	`minimax` or `qwen3`
`setup_mode`	enum	no	`voice_clone`	`voice_clone` or `voice_design`
`webhook_url`	string	no	—	Held — returns `503 webhook_delivery_unavailable`