Voice Cloning API — Clone a Voice for Speech or Music
Live2 modelsMuapi's voice cloning API covers two distinct products behind one unified REST endpoint: MiniMax Voice Clone for speaking-voice clones you plug into any MiniMax text-to-speech model, and Suno Custom Voice Cloning for singing-voice clones you plug into Suno music generation. Submit a short reference recording, get back a reusable voice ID, and use that ID in every future generation request — no per-request re-upload needed.
MiniMax Voice Clone
Clone a speaking voice from one reference audio clip. Reproduces tone, accent, rhythm, and emotion, then the resulting voice ID works with any MiniMax Turbo or HD speech model for text-to-speech generation.
Suno Custom Voice Cloning
Clone a singing voice in two takes: a 10-second sample plus a live readback of a system-generated phrase as an anti-deepfake liveness check. Returns a reusable voice_id for Suno music generation.
What is a Voice Cloning API?
A voice cloning API analyzes a short audio sample and produces a reusable voice model — a voice ID — that reproduces the speaker's tone, accent, rhythm, and delivery. Instead of re-uploading a reference clip on every generation, you create the clone once and reference the voice ID afterward.
Muapi exposes two voice cloning endpoints for two different downstream uses: MiniMax Voice Clone, whose output voice ID feeds any MiniMax Turbo or HD speech model for narration and dialogue, and Suno Custom Voice Cloning, whose output voice ID feeds Suno's music generation models as a persona_id for singing vocals.
Key Capabilities
Speech Voice Cloning
MiniMax Voice Clone reproduces a speaker's tone, emotion, accent, and rhythm from one reference clip, then generates new speech from any text input using that cloned voice.
Reusable Across Every Turbo/HD Model
A MiniMax cloned voice ID isn't locked to one model — use it with speech-02-hd, speech-02-turbo, speech-2.5, or speech-2.6 HD/Turbo interchangeably.
Singing Voice Cloning for Music
Suno Custom Voice Cloning captures a singing voice for use as a vocal persona in Suno music generation, remix, and extend requests.
Anti-Deepfake Liveness Check
Suno's cloning flow requires two takes: a 10-second sample, then a live readback of a system-generated random phrase, so a static recording alone can't be used to clone someone's voice without consent.
Noise Reduction & Normalization
MiniMax Voice Clone supports optional noise reduction and volume normalization on the input sample for cleaner clones from imperfect recordings.
Custom Voice IDs
Assign your own custom_voice_id (MiniMax) or voice_name/style tags (Suno) so cloned voices are easy to identify and reuse across projects.
Voice Cloning Model Comparison
| Model | Provider | Price | Use Case |
|---|---|---|---|
| MiniMax Voice Clone | MiniMax | $0.65 / clone | Speaking voice for text-to-speech narration and dialogue |
| Suno Custom Voice Cloning | Suno | Free (preview) | Singing voice persona for Suno music generation |
How to Clone a Voice via API
- Pick a target use case — MiniMax Voice Clone for speech/narration, Suno Custom Voice Cloning for singing vocals in music.
- Record a clean sample — one reference clip for MiniMax, or a 10-second sample plus a live readback for Suno's liveness check.
- Submit the request —
POST /api/v1/{model-slug}with the audio URL and any optional voice ID, style, or noise-reduction parameters. - Poll for completion — check
GET /api/v1/predictions/{request_id}/resultuntilstatusiscompletedand get back a voice ID. - Reuse the voice ID — pass it into any MiniMax Turbo/HD speech model, or as a Suno
persona_idin music generation requests.
Frequently Asked Questions
What is the Voice Cloning API?
Muapi's voice cloning API turns a short reference recording into a reusable voice ID. MiniMax Voice Clone covers speaking voices for text-to-speech; Suno Custom Voice Cloning covers singing voices for music generation.
Can I use a cloned voice with any text-to-speech model?
A MiniMax cloned voice ID works with any MiniMax Turbo or HD speech model — speech-02-hd, speech-02-turbo, speech-2.5-hd-preview, speech-2.5-turbo-preview, speech-2.6-hd, and speech-2.6-turbo.
How does Suno's anti-deepfake check work?
Suno Custom Voice Cloning requires two takes: a 10-second sample, then a live readback of a random phrase the system generates on the spot. This liveness check prevents cloning a voice from a static recording without the speaker present.
How much does voice cloning cost?
MiniMax Voice Clone costs $0.65 per clone. Suno Custom Voice Cloning is free during its preview period.
Can I clone a voice for singing rather than speech?
Yes — use Suno Custom Voice Cloning, which is purpose-built for singing vocals and integrates directly with Suno's music generation, remix, and extend endpoints.
Can I get Voice Cloning API access right now?
Yes. Sign up at muapi.ai, create an API key from your dashboard, and start cloning voices immediately — no waitlist required.