Google Omni AI Video Editor
Edit videos with Google Omni in Magic Hour. Describe the change you want — remove objects, replace backgrounds, add elements, restyle scenes, or enhance clips — and generate an edited result without timeline editing.
Spaceship to watch
View prompt
Replace the toy spaceship with a luxury wristwatch, keep the desk, lighting and camera move the same
Marble physics
View prompt
Make the marble roll off the edge of the table and bounce on the floor with realistic physics
Ferret character swap
View prompt
Replace the cat with a ferret doing the same actions
Museum style transfer
View prompt
Restyle the whole video as a watercolor painting
Outfit change
View prompt
Change her white dress into a deep red velvet gown, keep her, the violin, her playing and the room the same
City building reimagine
View prompt
Turn the building into a futuristic glass tower covered in hanging gardens
Glowing fern
View prompt
Make each leaf glow softly and release floating light particles where the hand touches it
Voxel transform
View prompt
Transform the whole scene into a voxel art world made of small cubes
Camera zoom reimagine
View prompt
Change the camera move to a slow zoom in toward her face
Trusted by teams at
How To Edit Videos with Google Omni
Three simple steps to edit AI videos with Google Omni in Magic Hour
- 1
Upload Your Video
Start with a clip you want to edit. Trim to 3–10 seconds — Google Omni works on short segments. - 2
Describe the Edit
Type what you want to remove, replace, add, change, restyle, or enhance. Optionally add up to 10 reference images for more control. - 3
Generate Your Edit
Click generate to apply the prompt. Once it's ready, download the edited clip or try another prompt.
Why Creators Choose Google Omni
Where Google Omni really shines: natural-language video edits, object and background changes, restyles, and short-clip cleanup without opening a timeline editor.
Edit with prompts
Describe the change in plain language — remove clutter, swap a background, or restyle a scene — and Google Omni generates the edited clip.
Reference images
Add up to 10 reference images when you need tighter control over what gets added or replaced.
Short-clip workflows
Built for 3–10 second edits — ideal for social clips, ad variations, and quick creative tests.
No Timeline Skills Required
Skip complex editors. Upload a clip, write a clear prompt, and iterate until the edit looks right.
Thousands of Templates
Make your photos even better and faster with our high-quality templates.
Explore TemplatesHow Google Omni Compares to Other Video Models
Google Omni THIS MODEL | Veo 3.1 | Kling 3.0 | MiniMax H3 | LTX-2.5 | Seedance 2.5 | |
|---|---|---|---|---|---|---|
Best for | Prompt-based video editing on short existing clips | Premium realism, polish, dialogue, and strong prompt adherence | Multi-shot storytelling, character references, structured cinematic control | Reference-driven clips with native audio, character consistency, longer short-form | Fast iteration, synced audio, uploaded audio, clips up to 60s | Cinematic continuity, structured references, narrative short-form |
Visual fidelity | Strong for edits | Premium realism | Strong | Strong | Good | Strong |
Real-human image to video | Supported | Better suited | Better suited | Better suited | Good with the right source image | Good |
Viral clips | Strong for edit variations | Good | Strong | Strong | Useful for rapid drafts | Strong |
Multi-scene storytelling | Not the primary use | Strong | Excellent | Medium to strong | Long clips; test continuity | Strong |
Surreal / imaginative concepts | Strong for restyles and creative edits | Strong | Good | Good | Prompt-led generation | Strong |
Native audio | No exposed generation setting | Yes | Yes | Yes | Yes | Yes |
Template fit | Good for short social and ad edits | Good | Good | Good | Good | Good |
What Google Omni is best at
Where Google Omni really shines: prompt-based edits on existing footage, object and background changes, restyles, and short social or ad variations.
Object removal and cleanup
Erase distractions, clutter, props, or unwanted details from a clip with a simple prompt.
Background and scene changes
Swap settings or restyle the environment without reshooting the original footage.
Add and replace elements
Insert products, props, or visual details — optionally guided by reference images.
Creative restyles
Transform the look of a clip into cinematic, stylized, or branded directions from text.
Fast short-clip iteration
Optimized for 3–10 second segments where you want quick edit variants, not long-form timelines.
Ad and social variations
A natural fit for testing creative directions on existing UGC, product, or campaign clips.
Google Omni — Model Card
Key specs, capabilities, and limitations.
Cost
48–384 credits/sec depending on resolution
Resolution
360p, 720p, 1080p, 4K
Aspect ratios
9:16, 16:9
Max duration
3–10s per edit
Audio
No native audio-generation setting in Magic Hour
Plans
Creator / Pro / Business
Limitation
Short clips only — up to 10s per edit or generation
Frequently asked questions
Google Omni is Magic Hour's name for Google's Gemini Omni Flash model. In AI Video Editor it edits existing videos from natural-language prompts — remove objects, replace backgrounds, add elements, restyle scenes, and enhance clips. It is also available in Text-to-Video and Image-to-Video for short new clips.
Google Omni is best for short prompt-based edits: cleanup, background swaps, object adds/removes, restyles, and creative variations of footage you already have.
Inside Magic Hour, Google Omni outputs 360p, 720p, 1080p, or 4K video at 24fps in 16:9 or 9:16, works on 3–10 second clips, costs 48–384 credits per second depending on resolution, and supports up to 10 optional reference images in AI Video Editor.
No. Google Omni has no audio option in AI Video Editor, Text-to-Video, or Image-to-Video.
Yes. Pick Google Omni in Text-to-Video or Image-to-Video to generate a new 16:9 or 9:16 clip. In AI Video Editor it edits a video you upload instead.
Google Omni costs 48–384 credits per second of output depending on resolution. At 720p it costs 144 credits per second (6 credits per frame at 24fps). Available on Creator, Pro, and Business plans.
No. Veo 3.1 is Google's premium video generation model. Google Omni (Gemini Omni Flash) is a separate model that edits existing clips from a prompt and generates short clips up to 10 seconds.
Each edit must be between 3 and 10 seconds.
Yes. You can attach up to 10 reference images to guide what gets added or replaced in the edit.
Magic Hour puts Google Omni next to generation models like Veo, Kling, MiniMax, and Seedance, so you can edit existing footage and generate new clips in one place.
Choose Google Omni for short prompt-based video edits or new clips up to 4K. Wan 3.0 and LTX-2.5 are also available in AI Video Editor. For generation with native audio, compare Seedance 2.5, Kling 3.0, MiniMax H3, Wan 3.0, or LTX-2.5 on your actual brief.
Paid subscriptions, credit-pack purchases, and separately agreed paid usage-based billing arrangements include commercial-use rights. Users without one of these paid arrangements are limited to personal, non-commercial use. Outputs made with Google Omni on Magic Hour can be used commercially, subject to our Terms.
Related Models
Compare models and create from text, images, or clips with Magic Hour's AI video generator.