Is MiniMax H3 Extender available online?
Yes. The dedicated H3 Max Extend task is connected to the generator on this page. Upload a source video, enter a continuation prompt, choose the output settings, and generate online.
Upload an existing video, describe what should happen next, and generate a longer continuation with the dedicated H3 Max Extend workflow. Choose the continuation length, output shape, and resolution before you create.
Upload a 1.625–60 second source video, then describe what should happen next.
Describe the continuation, not what already happens in the source video.
Preview video - Generate your own video above
Watch a source journey carry forward as the traveler and dogs move toward the village. The example shows the dedicated continuation workflow; results still depend on the source clip and the direction in your prompt.

MiniMax H3 Extender is an online video continuation tool built around the dedicated H3 Max Extend workflow. It reads an uploaded clip as context, follows a prompt for what happens next, and returns a longer downloadable video. The tool is available here now; it is generation, not timeline editing, so exact visual and audio continuity can vary.
Use the source video as temporal context, then make the continuation specific with a forward-looking prompt and delivery settings that match your next placement.
Upload the clip and write only the continuation: what the subject does, how the camera moves, and what changes after the current ending. This gives the model a clearer next beat for narrative, product, or movement-driven footage.

The source clip supplies the people, objects, camera movement, and scene leading into the new segment. A clear continuation prompt helps preserve the intended direction, while complex cuts, fast motion, and audio transitions may still produce a visible seam.

Set a 5–15 second continuation, select an automatic or explicit aspect ratio, and choose 480P, 768P, 1080P, or 2K output. The form shows the credit estimate before you submit.

The dedicated form keeps the source, continuation direction, and output choices together in one workflow.
Choose a supported video from 1.625 to 60 seconds, no larger than 50MB. Use a clip whose final subject and motion are easy to read.
Write the continuation rather than summarizing the existing footage. Name the next action, camera move, setting, and pacing that matter.
Choose 5–15 seconds, an aspect ratio, and a resolution up to 2K. Review the credit estimate, generate, then preview or download the extended video.
Add a purposeful next beat when a promising clip ends before the action, reveal, or transition is complete.
Carry a character into the next action or location without rebuilding the opening shot from scratch.
Extend a turn, camera move, or environmental reveal so the product reaches the intended final composition.
Generate an additional motion segment for landscapes, interiors, and background scenes before editing a loop.
Prompt the shot to keep pushing, tracking, orbiting, or pulling back from the motion already on screen.
Give a short vertical or square clip more time for a reaction, visual payoff, or clean final hold.
Explore the next blocking or camera beat in a previs sequence before committing to a production edit.
Key details about source files, continuation length, output controls, continuity, and repeated extensions.
Yes. The dedicated H3 Max Extend task is connected to the generator on this page. Upload a source video, enter a continuation prompt, choose the output settings, and generate online.
The uploaded source must be between 1.625 and 60 seconds. It must be 50MB or smaller and have an aspect ratio between 0.4 and 2.5.
You can choose a continuation from 5 to 15 seconds in one-second increments. The source length and chosen continuation are separate controls.
The uploader accepts MP4, MOV, WebM, and M4V files. The server also checks the file size, duration, aspect ratio, and presence of a readable video track before charging the task.
Choose 480P, 768P, 1080P, or 2K. Aspect-ratio options include Auto, 16:9, 9:16, 21:9, 4:3, 1:1, and 3:4.
Not always. The source provides context, but fast motion, abrupt source endings, complex cuts, or large prompt changes can create visible or audible seams. Use a clear ending and prompt a plausible next action.
Start with the very next action, then add camera movement, environment changes, pacing, and any sound direction that matters. Avoid spending the prompt retelling what viewers can already see.
You can download a completed result and use an eligible video as the source for another generation. Each pass is a new task, so continuity should be reviewed at every join.
Upload the footage you already have, direct the next 5–15 seconds, and generate a longer video in the dedicated extension workflow.