Voice Cloning

Create a reusable custom voice with sonus-tts from your own recordings.

POST https://api.sagea.space/v1/audio/clone

Authentication

Send Authorization: Bearer $SAGEA_API_KEY on every request. Get a key

Parameters

Send a multipart/form-data body with the following fields.

Parameter	Type	Required	Description
`name`	`string`	Yes	Human readable name for the cloned voice.
`audio_samples`	`file`	Yes	1 to 3 audio files, each 10 seconds or longer.
`language`	`string`	No	Default is `en-US`. Also supports `ne-NP`, `hi-IN`, and `es-ES`.
`consent`	`boolean`	No	Confirm you own the rights to the sample audio. Default is `false`.
`emotion`	`string`	No	Base style of `neutral`, `friendly`, `professional`, `excited`, `calm`, or `warm`.
`voice`	`string`	No	Base timbre hint of `nova` or `atlas` for shaping the clone.

Example request

Clone a Nepali-capable voice for Aarav Sharma from two samples.

curl -X POST https://api.sagea.space/v1/audio/clone \
  -H "Authorization: Bearer $SAGEA_API_KEY" \
  -F "name=Aarav Sharma voice" \
  -F "[email protected]" \
  -F "[email protected]" \
  -F "language=ne-NP"

Use clear recordings of 10 seconds or more per file for best results.

Response

A successful clone returns a reusable voice ID.

{
  "voice_id": "vox_aarav91np",
  "name": "Aarav Sharma voice",
  "language": "ne-NP",
  "status": "ready",
  "samples_received": 2,
  "model": "sonus-tts",
  "created_at": "2026-02-10T08:30:00Z"
}

Reuse voice_id in speech and streaming calls alongside nova and atlas.

Errors

Status	Code	Fix
`400`	`missing_artifact`	Attach 1 to 3 valid audio sample files.
`400`	`unsupported_format`	Upload `wav`, `mp3`, `ogg`, or `flac` samples only.
`401`	`unauthorized`	Re-export a valid `SAGEA_API_KEY` value.
`422`	`voice_not_found`	Check the returned `voice_id` starts with `vox_`.
`429`	`rate_limit_exceeded`	Cloning is limited daily, retry tomorrow.

Limits

Voice cloning allows 5 requests per day on Pro with custom limits on Enterprise.