01
Turn on Voice Clone
When you create a room with spoken translation, choose Voice Clone instead of Standard Voice, and confirm that your speakers agreed to have their voices cloned for this event.
Voice Clone
Live translation in each speaker's own voice.
With Voice Clone, your audience hears every speaker translated live in that speaker's own voice. Turn it on when you create a room. Captions and transcripts keep the same words, and there is no extra charge.
Free 15-minute test room, one per account
Definition
What is voice clone translation?
Voice clone translation is live speech translation where the translated audio sounds like the original speaker instead of a stock synthetic voice. A keynote given in English can be heard in Japanese, Thai, or Spanish in the speaker's own voice, while captions and transcripts carry the same translated words.
- Voice clone translation keeps the speaker's voice in the translated languages.
- Pikka Speech learns each voice during the event, with no recording session beforehand.
- Captions, transcripts, and room controls work the same way as with Standard Voice.
How it works
From the first word to their own voice
02
Speakers talk as usual
There is nothing to upload. While Pikka Speech learns a new speaker's voice from their first sentences, those sentences are spoken in Standard Voice.
03
Listeners hear the real voice
From then on, each translated sentence is spoken in the voice of the person who said it. Captions keep streaming as usual.
Panels & shared microphones
The right voice for every speaker
In a panel discussion the speaker changes every few sentences, often on one shared microphone. Voice Clone recognises who is talking by their voice and gives each person their own cloned voice.
- Each speaker is recognised by their voice, even when several people share one microphone.
- Translations are split where the speaker changes, so each sentence is spoken in the voice of the person who said it.
- When Pikka Speech is not sure who is talking, it uses Standard Voice rather than guess.
Consent & privacy
Consent first, deleted after
- Voice Clone only runs after the host confirms that the speakers agreed to have their voices cloned for the event.
- Each cloned voice is private to your event and is only used to speak that event's translations.
- Cloned voices are deleted when the event ends. If an event is interrupted, an automatic check removes any voice left behind, normally within 48 hours.
Comparison
Standard Voice vs Voice Clone
Both use the same live translation. The only difference is whose voice your listeners hear.
| Standard Voice | Voice Clone | |
|---|---|---|
| Sounds like the original speaker | ||
| Live captions and transcripts | ||
| Panels and shared microphones | One voice per speaker | |
| Voice samples to record | None | None, learned live |
| Speaker consent | Not needed | Required |
| Price | Included | Included, no extra charge |
- Choose Standard Voice when one consistent, neutral voice suits the event.
- Choose Voice Clone for keynotes, panels, and talks where the audience should hear the speaker, not a narrator.
Questions
Voice Clone, answered
Explore more
Language coverage
99 speaker languages and 142 listener languages and dialects.
Hybrid simultaneous interpretation
AI and human interpreters on the same language channels.
Webinar & livestream translation
Live translation for webinars and video streams, with QR join.
Event displays
Live captions on LED walls and projectors.
Last updated: 2026-09-24
Let your audience hear the speaker, not a narrator
Create a room, choose Voice Clone under spoken translation, and your speakers are heard in their own voice across your listener languages.