mimo-v2.5-tts-voiceclone (Voice Cloning Speech Synthesis)

Actual API Endpoint: POST https://api.tokenhot.ai/v1/chat/completions

Model Code: mimo-v2.5-tts-voiceclone

Clone any voice from an audio sample for speech synthesis. Simply provide an audio sample (Base64-encoded) to accurately clone the target voice and generate speech.

The audio sample is passed via the audio.voice field in the format data:{MIME_TYPE};base64,{BASE64_AUDIO}.

The Base64-encoded string must not exceed 10 MB, and only mp3 and wav formats are supported.

The text to synthesize goes in the role: assistant message; the role: user message is an optional style/tone instruction.

Style control via natural language or audio tags is supported (same as mimo-v2.5-tts).

Singing mode, preset voices, and voice design are not supported.

The low-latency streaming feature is not yet available; the streaming interface falls back to compatibility mode (returns all at once after inference completes).

Audio MIME Types:

Format	MIME_TYPE
MP3	audio/mpeg or audio/mp3
WAV	audio/wav

Bearer Token

Provide your bearer token in the

Authorization

header when making requests to protected resources.

Example:

Authorization: Bearer ********************

curl --location 'https://api.tokenhot.ai/v1/chat/completions' \ --header 'Authorization: Bearer <token>' \ --header 'Content-Type: application/json' \ --data '{ "model": "mimo-v2.5-tts-voiceclone", "messages": [ { "role": "user", "content": "" }, { "role": "assistant", "content": "Yes, I had a sandwich." } ], "audio": { "format": "wav", "voice": "data:audio/mpeg;base64,SUQz...(Base64-encoded audio sample)" } }'

{ "id": "chatcmpl-abc123def456", "object": "chat.completion", "created": 1774324467, "model": "mimo-v2.5-tts-voiceclone", "choices": [ { "index": 0, "message": { "role": "assistant", "audio": { "id": "3864a4690d8a4aaea5f6acf4853a0600", "data": "UklGR...(Base64-encoded audio data)", "expires_at": null, "transcript": null } }, "finish_reason": "stop" } ] }

mimo-v2.5-tts-voiceclone (Voice Cloning Speech Synthesis)

Request

Responses