Ship in Rough Waters
A ship sailing through rough waters, grey skies, cinematic ocean spray.
Text or image to video · 2K · Native audio
MiniMax H3 AI Video Generator
Create 4–15 second MiniMax H3 videos from text or one to two frame images. Describe the scene, motion, camera direction, dialogue, music, and ambience, then generate a fixed 2K video with native audio.
0 / 2000
Six MiniMax H3 workflow directions
Asteroids, Diamonds and Stars
Asteroids and diamonds and stars floating in the skies, 4k high quality, camera does not move.
Shocked and Surprised
The woman looks shocked and surprised, camera slowly zooms out.
Mountain Airglide
GoPro footage of someone airgliding through the mountains, shaky camera footage.
Dragon over New York
A dragon flying over New York City, drone shot.
Desert Poppy Bloom
A grassy surface, clear skies, rocky desert, millions of red poppy flowers slowly grow out of the grass.
Plan each input before choosing a mode
Treat every input as a production decision. Available input types and controls must be confirmed against the model version you can actually access.
Text direction
Describe the subject, visible action, camera, environment, lighting, pacing, sound, and final frame.
Image references
Choose references for identity, product appearance, composition, wardrobe, palette, or the opening frame.
Video references
Identify the exact motion, performance, camera path, edit rhythm, or environment behavior you want to guide.
Audio direction
Write dialogue, ambience, music, and sound cues separately, then confirm which controls the available mode supports.
Build a clearer multimodal brief
A useful H3 brief states what each reference controls, what may change, and what must remain stable.
Give every reference one jobLabel references by identity, style, motion, camera, sound, or composition instead of attaching files without direction.
Control the intended changeName the element to edit and list the subject, timing, lighting, framing, or environment details that should stay consistent.
Design one visible actionStart with one clear action and camera move so motion quality and prompt adherence are easier to review.
Choose text or frame guidanceStart from a text prompt, one opening image, or opening and closing frames, then describe the motion and sound that should connect the sequence.
Where this workflow can help
Use the planning structure for short-form concepts that need coordinated scene, motion, visual identity, and sound direction.
01
Trailer and story concepts
Outline scene beats, character continuity, camera movement, transitions, and sound before testing a short sequence.
02
Product and fashion ads
Lock the product, styling, lighting, talent, framing, and social format before exploring motion variants.
03
Character performance tests
Separate identity, expression, body motion, timing, and camera behavior so each reference has a clear role.
04
Targeted video edits
Define the background, dialogue, or scene detail to change while documenting the context that should remain unchanged.
Compare
Continue your MiniMax H3 video workflow
Generate directly on this page, or compare other connected video models when you need different input controls, output settings, or production speed.
MiniMax H3 FAQ
Current answers about MiniMax H3 generation, frame references, API modes, credits, and publishing checks.
Is MiniMax H3 available on ClipCanva?
Yes. Use the generator on this page for text-to-video or image-to-video creation through ClipCanva's connected KIE provider.
Which MiniMax H3 generation modes are connected?
ClipCanva currently connects KIE's MiniMax H3 text-to-video and image-to-video modes. Image-to-video accepts one opening image or opening and closing frames.
What should I prepare for a multimodal video brief?
Prepare the scene prompt, subject and product references, motion or camera references, sound direction, target format, and a list of details that must remain stable.
Can I use image, video, and audio references together?
The KIE API also documents a multimodal reference-to-video mode, but the ClipCanva generator currently supports text and up to two frame images. Video and audio reference uploads are not yet exposed in this interface.
Can I use MiniMax H3 output commercially?
Check the current KIE plan, MiniMax model terms, source-asset rights, and output rules before publishing. ClipCanva does not grant additional rights beyond the applicable provider terms.