Clone your own voice for the AI

On the Voice page in the dashboard you can clone your real voice. Once cloned, audio replies the AI sends sound like you.

Recording requirements:
- Duration: total 30 seconds to 2 minutes is enough.
- Format: MP3, M4A, WAV or OGG.
- Quality: clean room, no music, no background voices, no phone speaker. A simple phone microphone in a quiet room is fine.
- Language: speak in the language you want the cloned voice to use most.

How to do it:
1. Go to Voice → Clone voice.
2. Type a name for the voice ('My voice', 'Receptionist Anna', etc.).
3. Drag & drop one or more audio files. Multiple short clips work better than one long one.
4. Click 'Create voice'. Cloning takes a couple of minutes.
5. Once ready, pick the cloned voice from the voices list and turn ON 'Audio responses'.

Tips:
- Read varied content (a story, a price list, a sample reply): gives the AI more inflection range.
- Avoid 'um', 'er' filler: the model captures your natural pauses better with cleaner speech.
- You can re-clone later if you find a better sample: the new clone replaces the old one only if you choose it as active.