Skip to main content
Lipsync and Video Translation share one lip sync engine with Speed and Precision modes. Lipsync is the engine on its own: you bring the video and your own audio, and the engine redraws the mouth to match (dialogue replacement only). Video Translation generates translated audio first and optionally runs the engine after; see how they relate.

Quick Example

Both video and audio accept a publicly accessible HTTPS link ({ "type": "url", "url": "..." }) or a file uploaded via POST /v3/assets ({ "type": "asset_id", "asset_id": "..." }, see Upload Assets). The response returns a lipsync_id. Optional fields cover captions, music stripping, speech enhancement, watermarking, partial lipsync with start_time and end_time, keep_the_same_format (preserve source resolution and bitrate), fps_mode ("vfr", "cfr" or "passthrough"), folder_id, and callback_url / callback_id. enable_dynamic_duration (default true) lets output duration adjust to the new audio length. See the create reference for the full list.

Manage Lipsyncs

CLI Usage

The CLI wraps every call and can poll for you with --wait: