Character & IP consistency
Upload turnaround sheets or portraits so the same character stars across every new environment.
Upload up to 9 images, 3 motion videos and 3 audio tracks — blend character identity, product design, camera movement and musical timing into one coherent 2K clip.
With reference to video AI powered by MiniMax H3, you cite uploaded assets directly inside your prompt by their tile labels. Facial features, product packaging, wardrobe details, camera choreography and vocal rhythm stay locked from opening beat to final frame across commercial spots and narrative sequences.

Combine visual and audio ingredients without node graphs.
Upload up to 9 PNG/JPG/WebP images, 3 MP4/MOV clips (2–15s) and 3 MP3/WAV audio tracks.
Write how each asset participates — for example: subject from Image 1 inside Image 2, camera move from Video 1, pacing from Audio 1.
Keep Auto ratio or pick any of six ratios, choose 480p to 2K, and render your unified clip.
Go beyond single-frame animation.
Upload turnaround sheets or portraits so the same character stars across every new environment.
Pair a clean product packshot with a model portrait and studio background in one coherent ad.
Drive camera movement from a reference video while matching vocal or musical phrasing from an audio clip.
Up to 9 reference images, 3 reference videos and 3 reference audio clips in a single run (each clip 2–15 seconds, up to 15 seconds total per media type).
The first 5 reference images and all audio clips are free; images 6–9 add 1 credit each, and reference video seconds are billed at the same per-second rate as output video.
MiniMax H3 treats First/Last Frame and Omni Reference as two distinct modes — switch between them at the top-right of the studio.
Yes. Choose Auto to follow your references or lock 16:9, 9:16, 1:1, 4:3, 3:4 or 21:9.