Direct subject action and camera movement with MiniMax H3 motion control. Start from text or an image, then describe the direction, pace, framing, and ending of your shot. This generator uses MiniMax H3 with natural-language motion guidance; numeric camera trajectories are not available here.
Preview video - Generate your own video above
This existing H3 Max image-to-video output combines a rider’s forward motion with a tracking camera that gradually widens and rises. It illustrates prompt-directed movement, rather than a numeric camera trajectory or exact motion-transfer test.
Separate what moves in the scene from what the camera does. Text gives creative direction, while images establish appearance and composition. Results can vary between generations.
Describe who moves, what they do, their direction, and their pace. Choose one achievable action: walking toward a doorway, reaching for a cup, or riding steadily across a field.
Specify a push-in, pull-back, pan, tilt, orbit, or tracking move. State whether the subject stays centered and whether the camera distance should stay constant or change.
A first frame sets the opening view; a last frame guides the destination. Image references guide appearance. Multimodal references can guide movement, but do not provide a skeletal rig or guaranteed frame-for-frame motion transfer.
Build one clear shot, generate it, then adjust the movement you want to improve. The generator and the guidance below belong to the same workflow.
Use Text to Video for a new scene, or Frame to Video to animate an opening image. Use image or multimodal references when you need additional visual or motion guidance.
Write the subject’s action first, then the camera’s direction and framing. Add pace and a final composition. Keep the requested action achievable within your chosen duration.
Adapt a template below, paste it into the editable prompt, choose duration and resolution, and generate. For image inputs, keep the prompt consistent with the uploaded scene.
Check screen direction, camera movement, anatomy, and identity. Change one major instruction at a time. Keep a simpler take when a complex combined move produces drift.
Copy a template into the generator and replace the subject and setting. For image-to-video, describe the uploaded scene instead of introducing a different one. Use one main camera move per shot before combining moves.
Keep the viewpoint stable so the subject’s action reads clearly.
One continuous shot of an adult standing beside a desk in a daylight studio. The camera stays fixed in a medium-wide view. The person slowly reaches for a cup, lifts it to chest height, then holds still. Keep the face, clothing, cup, desk, and lighting consistent. Natural hand contact and balanced posture. Quiet room ambience. No cuts or camera movement.
Move the camera closer while the subject stays stationary.
One continuous shot of a ceramic vase on a wooden table beside a window. The vase and table remain still. The camera slowly dollies forward from a medium view to a close view, keeping the vase centered and revealing its surface texture. Preserve shape, color, lighting, and background. Soft room ambience. No cuts, object rotation, or sudden acceleration.
Move away from the subject to reveal more of the setting.
One continuous shot beginning close to an adult hiker standing still on a grassy ridge. The camera smoothly moves backward to reveal the valley and distant mountains, keeping the hiker near the center. The hiker holds the same pose while grass moves gently in the wind. Preserve identity and landscape. Natural wind ambience. No cuts or abrupt changes of scale.
Rotate the viewpoint from a fixed camera position. Use a pan for horizontal movement or a tilt for vertical movement.
One continuous landscape shot from a fixed tripod position. The camera slowly pans from left to right across a quiet harbor, ending on a lighthouse. Keep the horizon level and the camera position stationary; do not dolly or orbit. Gentle waves move naturally. Consistent daylight and geometry, distant seabirds and water sounds. No cuts or rapid turns.
Move around a stationary subject while keeping the camera distance similar.
One continuous shot of a small stone sculpture on a pedestal in a courtyard. The sculpture stays still. The camera makes a shallow clockwise arc around it from a front view toward a three-quarter view, keeping distance and height approximately constant. The background shifts with natural parallax. Preserve sculpture shape and courtyard geometry. Quiet outdoor ambience. No cuts or sculpture rotation.
Match a moving subject’s speed to hold a readable composition.
One continuous shot of an adult walking at a steady pace along a garden path. The camera tracks alongside at matching speed in a medium side view, keeping the person at a consistent size. The path and plants pass naturally in the background. Preserve face, clothing, walking direction, and lighting. Footsteps and gentle wind. No cuts, sudden zooms, or duplicated limbs.
Use the visible result to decide which input or instruction to change. More instructions do not always create more control.
State that the subject stays stationary and only the camera moves. Describe changing background parallax. Keep an orbit shallow before asking for a larger turn.
Use one unambiguous direction, a steady pace, and a clear endpoint. Avoid conflicting requests such as a locked camera and an orbit in the same shot.
Use a clear starting image with a readable pose. Reduce action complexity and camera travel, and repeat only the essential identity and object constraints.
H3 Max Camera Controls is a separate image-to-video capability for camera keyframes in 3D space. This page currently uses the ordinary MiniMax H3 generator; the controls described here are explanatory and are not form inputs.
Azimuth describes the camera’s horizontal angle around the scene. It is useful for planning an orbit rather than asking the object itself to rotate.
Elevation changes the viewing height or vertical angle, while distance changes how close the camera is to the scene. Together they describe a camera pose.
A sequence of timed poses describes a camera path. Writing these numbers into the current text prompt does not activate structured trajectory control; use natural-language directions in this generator.
Understand prompt guidance, references, camera movement, and the current online controls.
Yes, as natural-language direction for subject action and camera movement using the existing MiniMax H3 generator. This page does not provide numeric camera paths, a motion brush, or skeletal pose controls.
Yes. Choose Frame to Video, upload the opening image, and describe what should move and what should stay still. You can also add a last frame to guide the ending composition. Motion and endpoint similarity are not guaranteed to be exact.
The existing MiniMax H3 task supports video references in Multimodal Reference mode. Describe which movement the reference should guide and what appearance should stay consistent. Reference guidance does not guarantee exact choreography, pose transfer, or frame timing.
A pan rotates the camera horizontally from a fixed position. A tracking shot moves the camera through the scene, often alongside a moving subject. Make that distinction explicit to reduce conflicting movement instructions.
No. H3 Max Camera Controls uses structured camera poses with azimuth, elevation, distance, and timing. Those trajectory controls are not connected to this page’s generator; the current workflow uses editable text prompts.
Begin with one main action and one camera move, use a clear first frame when composition matters, and keep the requested travel modest. Review each output and refine one instruction at a time. Exact trajectories and identical results across runs are not guaranteed.
Describe one action, choose one camera move, and create your next shot with the generator above.
