🎤 LongCat-Video-Avatar 1.5

Audio-driven avatar video generation with Meituan's LongCat-Video-Avatar 1.5 (INT8 DiT + DMD2 8-step distilled LoRA).

Give it a portrait and a voice clip and it renders a lip-synced video. Prompts work best when they describe the character, the action and the scene. Audio with a single clean speaker works best.

⏳ Runs on a shared ZeroGPU slice. 2 s clips are the quickest way to try it; 5 s takes several times longer and uses more of your daily GPU quota.

Mode

Image-to-video anchors the identity to your portrait; text-to-video invents the person.

Clip length
Examples