HeyGen Voice Speech to Speech
Use it ↗Speech to speech: re-says a recording up to 5 minutes in a voice cloned from a clip up to 3 minutes, keeping words and timing. Best for re-voicing narration in 39 languages.
Re-says a speech recording (up to 5 minutes) in a voice cloned from a single-speaker reference clip (up to 3 minutes). The clone is made for that run only. Words and timing are kept: output length matched the source exactly in all our tests, in English, Spanish, Portuguese, French, German, Japanese, Italian and Korean. Language is set to the reference voice or detected; background noise cleaning of both recordings is optional and slower. Best for: re-voicing narration or dialogue with a new speaker, recasting a take, same-language voice swaps. Not for: changing the words or the language of the speech; a library voice picker (you must bring a reference clip). Instead: text to speech with 24 library voices, HeyGen Voice; a voice already cloned in ElevenLabs, ElevenLabs Voice Changer; translation with lip sync, HeyGen Video Translate Speed.