Upload one image and plot the camera's path keyframe by keyframe. H3 Max uses that trajectory to create a camera-led video while aiming to keep the scene stable and recognizable.
Click, drop, or paste a first frame
JPEG, PNG or WebP · Up to 10 MB
Sign in to generate videos
Closing Orbit
Tighten the frame as you circle, pulling attention onto the subject.
Features
The goal is a more directed clip: use a stable reference scene, guide the camera, and reuse the same trajectory settings across different products or frames.
Upload one frame and guide the camera around people, liquids, or objects caught mid-air. The model is designed to limit scene motion and create cleaner parallax, though fine details may still vary between generations.
Use Cases
One starting frame can become a more directed shot. Here are four practical ways creators and teams can use it.
Orbit a sneaker, a watch, or a coffee bag to generate a 360-style product video from a single photo. Use it for product concepts, ads, and launch teasers, and review generated details before publishing.
“Create a turntable-style view from one starting image.”
Start with a splash, a jump, or a pour caught in mid-air and guide the camera through the moment. The result can create a cinematic time-freeze effect for social content.
“Create a VFX-style look from one suspended moment.”
Arc around a building render or an interior frame to show depth, light, and layout. An architecture orbit turns a still visualization into a guided look around the design.
“A moving viewpoint can make depth and layout easier to explore.”
Circle an object, a setup, or a scene while you narrate. Teachers and course creators can add a visual walkthrough from a single reference image without filming from every angle.
“A moving viewpoint can complement a flat diagram.”
How It Works
Three steps from a still image to a directed camera move.
Drop in a JPEG, PNG, or WebP image. Scenes with real depth and a clear subject tend to produce more stable results.
Start from a trajectory preset, then adjust each keyframe's angle, height, distance, and time.
Choose a duration and resolution, generate, then preview the clip and download the mp4.
Why It Works
Prompt-only AI video workflows can leave much of the camera movement to the model. H3 Max Camera Control lets you supply a keyframed path while using your image as a stable visual reference.
FAQ
H3 Max Camera Control is an AI camera control video generator built on MiniMax's H3 Max model. You upload one image and plot a camera path with keyframes. The model then creates a camera-led video while aiming to keep the scene stable. It is made for product reveals, time-freeze clips, and directed camera moves from a single still.
Prompt-only AI video workflows can leave more camera movement decisions to the model. Here you provide a structured path with keyframes or presets. Results can still vary, but the added controls make the intended move easier to reproduce and can help subjects retain their shape, pose, and textures.
It is available to paid users and does not include a daily free quota. Clips start at 80 credits for a 5-second 480P video (16 credits per second at 480P, 25 at 768P), and the current cost appears on the generate button before submission.
Credits are reserved when you submit a generation. If a run is confirmed as failed, those reserved credits are released automatically and the failed run is not charged. Temporary network or status-check errors may need to be resolved before the final generation status is known.
Yes. You may use videos generated through 3D Camera Control in commercial projects, subject to applicable law and our Terms of Service. You are responsible for reviewing each output and making sure your source images and intended use do not infringe third-party copyrights, trademarks, or publicity rights.
No. Upload an image, click a preset, and generate. If you want finer control, every keyframe's angle, height, distance, and timing is editable in the timeline, but none of it requires a 3D or editing background.
Scenes with real depth, subjects caught mid-motion like splashes or jumps, and people in physically stable poses often produce a clearer freeze effect. Flat or blurry images tend to produce weaker parallax, so start with your sharpest frame.
Durations run 5, 8, 10, 12, or 15 seconds at 480P or 768P. For social posts and web previews, 480P may be sufficient; 768P provides more detail for product pages and presentations.
Start with one image, plot the camera path, and generate a clip that explores perspectives beyond the original frame.
