Voice cloning and voice provider settings

How to choose between OpenAI, MiniMax and Heygen voice providers in Translate Audio, and when voice cloning makes sense.

Written By Umakhan Magomedov

Last updated 4 days ago

Translate Audio gives you a choice of three voice providers for the translated voiceover. The right choice depends on how closely the output needs to match the original voice and your token budget.

Voice cloning toggle

The Voice cloning toggle controls whether the translation uses the voice from the original audio or a standard synthetic voice.

  • On: the original speaker's voice is cloned and used for the translated output. The voice character, timbre and style carry over to the translation.

  • Off: a standard synthetic voice is used. Faster and cheaper, but the output does not sound like the original speaker.

Voice providers

When voice cloning is enabled, you can choose the provider:

Default (OpenAI)

The simplest and fastest option. Uses OpenAI's synthesis engine with a standard voice. No voice cloning is applied — the output uses a neutral AI voice. Best for quick translations where voice similarity is not important.

Voice cloning (MiniMax)

Clones the voice from the audio file using MiniMax. The translated speech sounds closer to the original speaker. Two options are available:

  • Clone from file: the voice is cloned automatically from the uploaded audio. Each new file triggers a fresh clone.

  • My MiniMax voice: use a saved Custom Voice from your account. Requires a MiniMax voice created in Custom Voices.

ℹ️ MiniMax cloning charges a one-time activation fee of 150 tokens the first time a specific voice is used. After activation, only the per-second translation cost applies.

Voice cloning (Heygen)

The highest-quality option. Heygen produces the most natural-sounding voice clone but costs more per second. Best for professional content where voice authenticity is critical.

Provider comparison

Provider

Voice similarity

Speed

Cost

Default (OpenAI)

Low (standard voice)

Fast

Lowest

MiniMax

Medium-high

Medium

Medium

Heygen

Highest

Slower

Highest

See Token pricing for each tool for exact costs per second.

How to change the provider

  1. Open Translate Audio and upload or record your file.

  2. Tap the Settings icon before tapping Translate.

  3. Toggle Voice cloning on or off, then tap Voice provider to select a provider.

  4. The token estimate updates immediately to reflect your choice.

Frequently asked questions

Related articles