Text, images, video and audio read as one context. Native 2K at 24 fps, stereo sound generated with the picture, and edits you ask for in a sentence. Available anonymously on Venice.
Try MiniMax H3 free and anonymously.
Three ways in
Venice opens all three H3 modes. Which one you choose decides what you need to prepare and which settings your assets take out of your hands.
Prompt only, no upload. The fastest way to try an idea out.
Give it the opening and the closing frame; it fills in the transition.
Steer the shot with your own material — a face, a motion, a voice.
Specs
Benefits
Words, pictures, clips and sound files are read together rather than one at a time. A shot can inherit a face from a still, a move from a clip and a voice from a recording in the same run.
Video arrives on a 1440-pixel short edge, roughly 3.7 megapixels at 21:9. It holds up on a client monitor, a TV or a store screen without a second render.
Dialogue, effects and ambience come back with the picture, lined up with what happens on screen. No separate scoring stage and no syncing afterwards.
Name the element that should change and the rest of the frame holds. A take you already like survives the revision instead of being rolled again.
Face, styling, motion and voice carry across shots when the same reference set is supplied each time. A reference recording can also give a character a different voice, or replace a line said on camera.
Your prompts, references and the recordings you feed it are never stored or trained on. Generated on Venice, and gone when you are done.
Anonymous on Venice.