Timed Multi-Shot Prompts
List each shot with its start time, such as 0-4s for the wide shot and 4-9s for the close-up. H3 cuts in that order inside one clip and keeps the characters the same across cuts.
Turn a prompt, a photo or a set of references into a 4-15 second video with native stereo sound in 768P or 2K, using the MiniMax H3 AI video generator.
The MiniMax H3 AI video generator renders picture and stereo sound in one pass, and it can cut between several shots inside a single 4-15 second clip.

Trailer Shots from One Sentence A one-sentence brief for a space-opera teaser became a wide shot of the bridge, a close-up of the captain and the empty window after the fleet jumps away.

Vertical Fashion Campaigns Close-ups, full-length looks and a closing brand card cut together in a 9:16 frame, ready for Reels and paid social.

Weight and Momentum in Action Snow sprays, the rider lands with real weight and the chase camera keeps pace, so fast action reads as one continuous run.

Dialogue for Vertical Drama Close coverage and reaction shots come back with the spoken lines already in them, so a short-drama beat needs no dubbing pass.

Animated Posters A flat layout comes alive as the character jumps out of the frame, while the headline, series number and small print stay sharp.
Write one brief across text, images, video and audio, and the MiniMax H3 AI video generator returns a 2K clip with stereo sound that follows it. Faces, camera moves and voices can each come from a file you name.




Bring the MiniMax H3 AI video generator into ads, fashion films, trailers and social clips. Sound comes back with every clip, so a first draft is already close to a finished cut.

Upload a product photo as the first frame, then describe a slow turn, a cut to a detail and a pull back out. The product keeps its shape and materials from every angle, ready for product pages and paid campaigns.

Add a portrait as a reference image and ask for a campaign with quick cuts built around her. Face, makeup and styling carry through each shot, and the pacing fits a lookbook or a launch teaser.

Write the tunnel, the walkout and the ring as separate shots with their own camera moves. H3 cuts between them inside one clip and keeps the fighter's face and robe the same, useful for pitch reels and teasers.

Upload a style frame or describe the look, such as hand-drawn line art with flat color. The style holds through every cut, and a title card can close the episode.

Describe a start menu, a loadout screen and the jump into gameplay. Buttons, cursor moves and HUD elements stay readable from frame to frame, ready for store pages, pitches and team reviews.

Describe the texture, the pour and the sound, such as honey over a split croissant and coffee filling a cup. Slow macro shots with matching sound effects suit social posts and café menus.
The MiniMax H3 AI video generator lets you direct shots, frames, references and sound in the same prompt.
List each shot with its start time, such as 0-4s for the wide shot and 4-9s for the close-up. H3 cuts in that order inside one clip and keeps the characters the same across cuts.
Upload an opening image, a closing image or both, and describe the path between them. H3 fills in the motion, so a product lands exactly on the pack shot you designed.
Combine 9 images, 3 videos and 3 audio clips in one generation and say what each one controls, such as identity, setting or style. When two references disagree, state which one wins.
Add a clip of a dance, a stunt or a camera move and point to it in your prompt. H3 transfers the movement to your own character and scene.
Add a voice sample, say which character speaks with it and write each line exactly as it should be heard. Dialogue is stable in 11 languages, including English, Chinese, Japanese, Korean and Spanish.
Choose any length from 4 to 15 seconds in 768P or 2K. Pick 16:9 or 21:9 for widescreen, 9:16 for TikTok and Reels, or 1:1, 4:3 and 3:4 for feeds.
The MiniMax H3 AI video generator follows direction on camera, performance and sound, from a single line of text to a scene built on reference clips.

Image to video: family dinner
First frame: a photo of a family dinner with a bowl of ramen in front. Pull focus to the people in the background and add more steam to the ramen bowl.

Voice from a reference clip
Reference video 1 and reference audio 1: the man in the pink suit says, "Follow the wind, live free. Leave worries behind, enjoy the moment." His voice follows the timbre of audio 1.

Felt pigeons in the city
Stop-motion look: a flock of knitted felt pigeons pecks at a slice of toast on a wooden bench in a busy European square. Low camera at bench height, pedestrians soft in the background, cooing, footsteps and street chatter.

Breaking under the overpass
Handheld documentary shot beneath a highway overpass: a breakdancer drops into windmills and a freeze while a crowd in dark hoodies watches in a circle. A boom-bap beat from a speaker, cheers on the freeze.

Papercut kite flight
Papercut animation: a small boy stands in tall paper grass under rolling orange hills at dusk, holding a kite string. The camera pulls back and rises to reveal a bright kite high in the deep blue sky. Soft wind and rustling paper.

Chasing the map
Two hikers on a misty moor lose their map to a gust of wind. Wide shot as they chase it downhill, then a low close-up as they dive into the heather, laughing, and pin it down. Wind and breathless laughter.
Generate video and sound in one pass

Step 01
Upload a photo to animate, or add images, video and audio as references.

Step 02
Write the shots, camera moves and sound. Put spoken lines in quotes.

Step 03
Set the length, resolution and aspect ratio, then generate and download.

Start with a line of text, a photo or a set of references. Generate a clip with sound, compare takes and download the one you want in 768P or 2K.
Upgrade or cancel anytime. Choose a plan for clean output, fast generation, full model access and priority support.
Monthly
Annual
Quick answers about MiniMax H3, plus the plans, credits and billing behind it on ImgVid.