Clone voice

Create a workspace voice from audio samples with ElevenLabs, Fish Audio or MiniMax instant cloning.

POSThttps://api.tried.com.au/clone-voice

Needs keys Creates a voice from one or more audio samples and adds it to your workspace's voices. Real cloning needs the chosen vendor's key (ELEVENLABS_API_KEY, FISH_API_KEY or MINIMAX_API_KEY).

Without that key, or with TRIED_MOCK_PROVIDERS=1, Tried stores the voice with a mock_ provider ID so you can build and test agents end to end.

Request#

Send multipart/form-data with the fields below and one or more audio file parts. A JSON body is also accepted (see JSON form).

Field	Type	Required	Description
`voice_name`	string	Yes	Name of the new voice. 1 to 100 characters.
`description`	string	No	Up to 500 characters. Sent to ElevenLabs and Fish Audio.
`provider`	string	No	`elevenlabs` (default), `fishaudio` or `minimax`.
`files`	file	Yes	Audio samples. Send at least one. All samples together can be at most 25 MB. MiniMax uses only the first file.

JSON form#

For scripts, send application/json with the same fields and the samples as base64:

{
  "voice_name": "Front desk (Sydney)",
  "provider": "elevenlabs",
  "files": [
    { "filename": "sample.mp3", "mimetype": "audio/mpeg", "content_base64": "SUQzBAAAAAAA..." }
  ]
}

filename defaults to sample.mp3 and mimetype to audio/mpeg. The whole JSON body is limited to 25 MB.

Example request#

curl -X POST https://api.tried.com.au/clone-voice \
  -H "Authorization: Bearer $TRIED_API_KEY" \
  -F voice_name="Front desk (Sydney)" \
  -F provider=elevenlabs \
  -F files=@sample1.mp3 \
  -F files=@sample2.mp3

import { readFile } from "node:fs/promises";

const form = new FormData();
form.append("voice_name", "Front desk (Sydney)");
form.append("provider", "elevenlabs");
form.append("files", new Blob([await readFile("sample1.mp3")], { type: "audio/mpeg" }), "sample1.mp3");

const res = await fetch("https://api.tried.com.au/clone-voice", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.TRIED_API_KEY}` },
  body: form,
});
const voice = await res.json();

import os, requests

with open("sample1.mp3", "rb") as f:
    res = requests.post(
        "https://api.tried.com.au/clone-voice",
        headers={"Authorization": f"Bearer {os.environ['TRIED_API_KEY']}"},
        data={"voice_name": "Front desk (Sydney)", "provider": "elevenlabs"},
        files=[("files", ("sample1.mp3", f, "audio/mpeg"))],
    )
voice = res.json()

Response#

201 Created

{
  "voice_id": "custom_voice_137bcfa732ab6e2df5d2b815",
  "voice_name": "Front desk (Sydney)",
  "provider": "elevenlabs",
  "provider_voice_id": "mock_9f060d81a4d8df9b04ca",
  "gender": "female",
  "accent": "Custom",
  "age": "middle_aged",
  "preview_audio_url": null
}

Field	Type	Description
`voice_id`	string	Starts with `custom_voice_`. Set it as an agent's `voice_id`.
`provider_voice_id`	string	The vendor's voice ID. In mock mode it starts with `mock_`. For MiniMax it is an ID Tried picks, starting with `Tried`.
`gender`	string	Always `female` for cloned voices.
`accent`	string	Always `Custom`.
`age`	string	Always `middle_aged`.
`preview_audio_url`	null	Always `null`.

Errors#

Status	When
`400`	A field fails validation (for example `voice_name` missing), `At least one audio sample (files) is required`, `Audio samples are limited to 25 MB in total`, or the vendor rejects the clone (`ElevenLabs voice clone failed (<status>): ...`, and the same for `Fish Audio` and `MiniMax`).
`401`	Missing or invalid API key.
`403`	The key's IP allowlist or scopes reject the request (this route needs the `agents` scope), or the caller is a read-only dashboard member.
`413`	A single file is over 50 MB, or the JSON body is over 25 MB.
`429`	Rate limit exceeded. See Rate limits.

See Errors for the body format.

Retell compatibility#

Same path and method as Retell, and both take multipart voice_name plus files. The provider field differs: Tried reads provider with values elevenlabs, fishaudio and minimax, and defaults to elevenlabs.