Input Image
3D Camera Control
Drag handles: ● Azimuth ● Elevation ● Distance
API Parameters
Output
Generated image will appear here
Drag handles: ● Azimuth ● Elevation ● Distance
Generated image will appear here
Hover sphere to preview, scroll to set zoom, click to add waypoint (2-4 points). Zoom: 5.0
Click on the sphere above to add camera positions
Generated keyframes will appear here
Prompts are sent as Next Scene: Describe the camera or scene changes you want.
Recommended range: 0.7–0.8
Generated next scene will appear here
Fixed prompt is applied server-side using the model README instruction.
LoRA path is managed on backend (env override supported).
Generated light transfer will appear here
Trigger word is fixed to 重新照明. Your instruction can be in any language.
LoRA path is managed on backend (env override supported).
Generated relight output will appear here
Generate video from the keyframes you created in Multi-image.
Generated video will appear here
Quick Stitch is instant. AI Video uses Seedance (queued).