Extraordinary scenes. Unforgettable moments. See what AI can bring to life on VideoAny.
Explore more with AIPreview page for MiniMax H3. Currently executes through Grok Imagine Video 1.5 in text-image-to-video mode.
Generate video · 6 credits total
This route is a preview of MiniMax H3. The current form runs Grok Imagine Video 1.5. Write a text prompt describing one visible subject, one action, and one camera movement. A video result is generated.
Describe one subject, one action with a clear start and end, and one camera move. Keep the scene simple and avoid multiple actions.
The model creates a short video from your prompt. No audio, dialogue, or music is generated. Review the visual action only.
Check that the action follows the described sequence. Verify the subject, setting, and camera move match your prompt.
A person sits by a rain-streaked window in a dim room, holding a warm mug. They take a sip, then look out at the rain. The camera slowly zooms in on their reflective expression.
Confirm the sip action occurs, then the look, and the zoom ends on the expression.
Add timing words like 'first... then...' to define the order. Example: 'First they take a sip, then they look out at the rain.'
Describe the final framing explicitly: 'ends with a tight close-up on the eyes' or 'zoom stops when the face fills the frame'.
State that the mug is held throughout the action: 'they hold the mug the entire time, even while looking out the window.'
