via “voice cloning and speech synthesis with mouth movement regeneration”
AI video/podcast editor — edit video by editing text, filler removal, eye contact, studio sound.
Unique: Combines speaker embedding (voice cloning) with video generation (mouth movement synthesis) in a single workflow — when user edits transcript text, the system regenerates both audio (cloned voice speaking new text) and video (mouth movements matching new speech). This requires tight coupling between speech synthesis and video generation models.
vs others: Integrated into text-based editing workflow (edit transcript → voice regenerates automatically) vs. standalone voice cloning tools (ElevenLabs, Descript's own AI Speech); but voice clones are locked to Descript platform, unlike ElevenLabs which provides API access.