A normal image-to-video workflow gives AI one visual starting point and lets it decide where the scene goes next. MiniMax H3 First and Last Frame gives you more control by defining both ends of the shot. Instead of asking AI to invent the entire video, you provide two visual anchors and direct the movement between them.
Control How Your MiniMax H3 Video Starts and Ends
Upload one image for the opening composition and another for the final composition. Your prompt describes the path between them: what the subject does, how objects move, how the camera changes, how lighting develops, and what should be heard.
MiniMax H3 can work with a first frame, a last frame, both images, or no image depending on the selected generation workflow. Its hosted workflow supports short videos at up to 2K resolution, with available durations varying by interface.
What Is MiniMax H3 First and Last Frame?
MiniMax H3 First and Last Frame is a keyframe-controlled AI video workflow that uses an opening image and an ending image to guide a generated video.
- The first frame establishes the opening subject, composition, camera angle, environment, product position, character pose, or visual style.
- The last frame establishes where those elements should arrive by the end of the clip.
- The prompt explains the motion and changes between those two states.
The model treats the task as a continuous path from the first image toward the final image. A good first and last frame AI video is not simply a morph between two pictures: it needs a plausible relationship between subject, camera, environment, lighting, and objects.
How MiniMax H3 First and Last Frame Works
- 01
Upload your first frame
Choose the image that defines the beginning of the video. It can be a product photo, character design, storyboard frame, fashion image, landscape, interior, vehicle, illustration, campaign visual, or generated image. Use a clear image with an intentional composition so the subject and details that need to continue are obvious.
- 02
Add your last frame
Upload the image that represents the intended ending. It can show a new pose, camera position, product state, lighting condition, environment, object arrangement, or completed transformation. The model now has a destination instead of having to invent an unpredictable ending.
- 03
Describe the motion between both frames
Describe what changes over time: subject movement, camera movement, pacing, environmental motion, sound, and how the scene approaches the final composition. Do not spend the whole prompt repeating visual details already supplied by the two keyframes.
- 04
Generate and review
Choose the available duration, aspect ratio, and resolution, then review subject consistency, object shape, motion direction, camera movement, transition timing, background stability, native audio, and final-frame accuracy. When refining, change one instruction at a time.
First and Last Frame AI Video Gives You More Control
Define the opening composition
The first frame gives MiniMax H3 an exact visual starting point. This is useful when you already have an approved product image, concept frame, campaign creative, character design, or storyboard.
Control the final moment
The second image gives the generated shot somewhere specific to land, which is useful for product pack shots, character pose changes, before-and-after scenes, camera reveals, transformations, and storyboard transitions.
Direct camera movement
Add one clear camera instruction when it helps connect the compositions: slow push-in, pullback, pan left, pan right, tracking shot, orbit, tilt, or static camera.
Generate video with native audio
Describe dialogue, environmental ambience, physical sound effects, or music alongside the visual transition. Sound direction belongs in the same shot brief as movement and camera behavior.
MiniMax H3 First and Last Frame Prompt Formula
Subject action + transition path + camera movement + environmental change + ending condition + audio
Product reveal
A closed silver watch case slowly rotates toward the camera as the lid opens. The camera makes a controlled push-in while narrow studio lights move across the brushed metal. The watch rises smoothly into the final upright position shown in the end frame. Dark studio background remains unchanged. Soft mechanical clicks, subtle room tone, no dialogue.
Character movement
The woman turns slowly away from the window and walks toward the center of the room. The camera tracks backward with restrained movement while her clothing and hairstyle remain consistent. Warm afternoon light gradually shifts into the cooler evening lighting shown in the last frame. Quiet footsteps on wood and distant city ambience.
Scene transformation
The empty architectural interior gradually becomes furnished as sunlight moves across the room. Chairs, tables, plants, and lighting appear through natural physical transitions while the camera remains nearly static. The composition converges on the completed interior shown in the final frame. Soft room ambience with subtle furniture movement.
Do not just repeat “start with Image A, end with Image B.” The images already establish that. Use the prompt to explain the visual path between them. For more structures, camera vocabulary, and examples, visit the MiniMax H3 Prompt Guide.
Best Uses for MiniMax H3 First and Last Frame Video
Product reveals
Start with a closed box, covered product, distant composition, or neutral pack shot. Use the last frame for the intended hero composition, then describe the reveal in between.
Before-and-after videos
Define an initial state and a finished result for room redesigns, fashion changes, product transformations, lighting changes, seasonal transitions, restoration concepts, and design iterations.
Storyboard to video
Turn two approved storyboard panels or keyframes into a moving shot. The prompt explains the action connecting the beginning and the next visual beat.
Character pose transitions
Move clearly between standing and sitting, looking away and looking at camera, walking and stopping, or reaching and holding an object. Keep identity, wardrobe, and composition compatible where they should remain consistent.
Camera transitions
Use two compositions of the same scene for a wide-to-close shot, front-to-side view, eye-level-to-overhead move, or an exterior-to-doorway reveal. Tell the model how the camera should travel between them.
Social and advertising creative
Use an opening frame that communicates the subject immediately and an ending frame built around a CTA, product, logo, or key visual. The generated motion connects two planned moments.
First & Last Frame vs Image to Video
Both workflows begin with visual inputs, but they solve different problems.
| Workflow | Visual control | Best for |
|---|---|---|
| First & Last Frame | Defines both the beginning and ending | Controlled transitions, transformations, storyboard shots, and product reveals |
| Image to Video | Defines mainly the starting visual | Animating one photo or letting the ending develop more freely |
| Reference to Video | Guides identity, style, motion, camera, voice, or other characteristics | Character consistency, product consistency, motion reference, and multimodal direction |
If you only have one source image and do not need an exact ending, use MiniMax H3 Image to Video. If you need character, product, motion, video, or audio references rather than a fixed final frame, use MiniMax H3 Reference to Video.
Tips for Better MiniMax H3 Start and End Frame Videos
Make the two frames visually compatible
The model needs to construct a believable path. When possible, keep important elements recognizable between frames rather than combining unrelated locations, subjects, or visual styles in one short transition.
Describe changes, not static images
Your keyframes already convey appearance and composition. Use the prompt for movement, timing, interaction, camera behavior, lighting, environmental motion, and sound.
Use one main camera move
For many shots, one deliberate push, pull, pan, orbit, track, or static composition is more controllable than several conflicting movements.
Explain the ending
Even with the final image supplied, state how the action should settle: for example, that the subject slows into the final pose or the camera completes its orbit into the final product composition.
Keep the transition physically plausible
If an object changes position, a person moves, or light shifts from day to night, explain the intermediate action that connects the two states.
MiniMax H3 First and Last Frame FAQ
Does MiniMax H3 support first and last frame video generation?
MiniMax H3 supports first/last-frame generation as an image-driven workflow. Depending on the selected interface and workflow, you can provide a first frame, a last frame, both images, or no image.
What is the difference between first frame and last frame AI video?
The first frame defines how the generated video begins. The last frame defines the intended final visual state. With both supplied, the model generates the motion and audiovisual path that connects them.
Can I turn two images into a video with MiniMax H3?
Yes. Upload one image as the first frame and another as the last frame, then describe how the scene should evolve between them. This gives you control over both anchor compositions rather than only animating one starting image.
Do my first and last frames need to be identical in size?
Compatible dimensions, framing, and aspect ratios make the workflow easier to control. Use images that fit the aspect ratio you plan to generate whenever possible.
What should I write in a MiniMax H3 first last frame prompt?
Focus on the path: subject movement, intermediate actions, camera behavior, lighting or environment changes, the final landing, and sound. The two images already communicate the visual endpoints.
Can MiniMax H3 generate sound with a first and last frame video?
MiniMax H3 is designed as an audiovisual workflow, so prompts can include dialogue, ambience, effects, and music direction alongside the visual transition.
Can I use only a last frame in MiniMax H3?
Availability depends on the selected workflow, but first/last-frame controls can accept one or two image conditions. Use the relevant image field in the image-to-video workflow.
Is first and last frame better than normal image to video?
It depends on how much control you need. Standard image-to-video is useful when you have a starting image and are comfortable letting the ending develop more freely. First-and-last-frame generation is better when the final composition is planned.
How long can MiniMax H3 first and last frame videos be?
MiniMax H3 supports short-form video generation. Available duration choices can vary between the model API and the current browser interface.