Voice Cloning
Create a custom AI voice from an audio sample
Voice Cloning lets you create a custom AI voice from a short audio sample. Upload a recording or use your microphone, then generate speech in that voice.
Requirements for your audio sample:
- Duration: 10 seconds to 5 minutes
- Formats: MP3, WAV, OGG, M4A, or WebM
- Max file size: 4.5 MB
How to Clone a Voice
Either upload a file by dragging it into the upload area, or click the microphone icon to record directly in your browser.
Choose the language for your cloned voice from the dropdown.
Type the text you want the cloned voice to speak. Maximum 500 characters for English, 300 for other languages.
Check the consent box, then click Generate or press Cmd+Enter (Mac) / Ctrl+Enter (Windows).
Supported Languages
Tips for Best Results
- Use clear audio — Minimize background noise and ensure the speaker is easy to hear
- Single speaker only — The sample should contain just one person speaking
- Natural speech — Conversational speech works better than reading or acting
- Longer is better — Samples closer to 5 minutes produce higher quality clones
Credits
Credits are calculated based on the length of text you generate. Cloned voice generation uses more credits than standard voices.
Check your credit balance at the top of the page before generating.