Lipsync API

Lipsync API uses talking-head AI models to synchronize mouth movements in a video or static photo with any audio track. The model analyzes the audio waveform, maps phonemes to mouth shapes, and renders natural lip movements frame by frame. It works with any language -- the synchronization is phoneme-based, not language-specific. You can input either a video (for re-dubbing) or a single photo (to create a talking head from a still image). Developers building dubbing pipelines, avatar applications, and content localization tools integrate lipsync as a core capability. A video platform can offer automatic dubbing in new languages. A chatbot app can make an avatar respond with natural lip movements. E-learning platforms can create talking-head instructors from a single photo. The API handles the ML complexity so you can focus on your product logic. For video input, ensure the face is clearly visible and front-facing for at least 80% of frames. For photo input, use a high-resolution frontal portrait. Audio should be clean speech without heavy background music. The API returns an async response for video processing -- use the webhook or polling endpoint to retrieve the result. Pricing is per second of output video.