Upload your image
Drop a photo, illustration, 3D render, or product shot. PNG, JPG, or WebP — any common format works. You can upload one or more reference images.
Image to Video with transparent model disclosure
Upload a photo, illustration or render and describe the motion. Gemini Omni Flash animates the still while preserving the intended subject and composition.
Reference image · Motion control · Flexible aspect ratios
● Video Generator
Animate your image
Upload a still, describe the motion and generate a cinematic clip with Gemini Omni Flash.
Image-to-video generator
0000 / 5000
● How it works
How image-to-video works on Gomni
From still to motion in a few steps. The model aims to preserve the reference, but identity, lighting and composition should be checked in every result.
Describe the motion
Tell the model what should move and how. "Camera pans left," "the subject turns toward the camera," "wind moves the hair" — natural language is enough.
Generate the clip
Gemini Omni Flash animates the still from your reference. Review identity, lighting and composition before publishing.
Refine or export
Review the result, revise the prompt in a new job if needed, and verify provider-specific watermark and usage terms before publishing.
● Benefits
Why image-to-video on Gomni works for real production
Reference fidelity is a result to measure, not a guarantee. Compare the first output frame with the source before using a clip.
Check faces, products, lighting and geometry for drift. Gomni has not published a controlled consistency score for this workflow.
Image-to-video features on Gomni
What you get when you animate stills through Gomni.
Multi-image reference
The current form supports a starting image and text direction. Multi-image, audio and video reference inputs are outside this Sprint's implemented scope.
Motion control via prompt
Describe motion in your own words. Pans, tracks, push-ins, subject actions, environment movement — the model understands camera language and physical motion.
Aspect ratio independent
Animate any input ratio into any output ratio. Vertical product shot to horizontal hero clip, or square reference to portrait Reel — no awkward cropping.
Consistent subject across clips
Generate multiple clips from the same reference and keep the subject consistent. Useful for character animation and product campaigns.
Conversational refinement
After the first generation, reply in natural language to adjust. "Slower motion," "camera moves the other way," "keep everything but add a subtle wind" — the model preserves what works and changes what you asked.
Up to 10-second clips
Same length cap as text-to-video. Long enough for product showcases, social ads, character beats, and reference-to-motion design.
● FAQ
Image-to-video FAQ
Common questions about animating stills on Gomni.
Animate your first image
Upload a still and generate with Gemini Omni Flash without a monthly commitment.