Conversational multi-turn editing
Nano Banana for video: describe the next change in plain language and the model applies it to the clip you already have, so a sequence of small corrections replaces a round trip through a timeline.
Google DeepMind
Gemini Omni
The conversational video editor. Edit video by talking to it. Gemini Omni takes text, images, audio, video or a rough sketch and returns a clip with native audio, then keeps refining it turn after turn while people, objects and scenes stay locked.
- Provider
- Google DeepMind
- Audio
- Native
- Editing
- Conversational, multi-turn
- Inputs
- Text, images, audio, video, sketches
- Open weights
- No
- Commercial use
- Allowed
Features
What Gemini Omni does differently.
Consistency lock
People, objects and scenes hold their identity across turns. A face, a product or a set stays the same from edit to edit rather than drifting or warping.
Cinematic by default
Orbital dollies, film-grade lighting and native audio come out of the box, so a plain prompt reads as a shot rather than a test render.
Any-to-any input
Text, images, audio, video and sketches all go in the same door and come back as video with sound, which makes a marked-up screenshot or a product photo a valid starting point.
Examples
Prompts that show its range.
Screenshot of map with sketched route, turn it into cinematic drone footage following that exact path at golden hour, don't show the drawing
Try this promptReplace the pomegranates with apples
Try this promptKeep me walking in the same shot, change outfit every second: streetwear, red dress, yellow trench, no warping
Try this promptOne continuous hyperlapse through ten world cities morphing on the beat
Try this promptPull back from the eyes of a Renaissance portrait, let her breathe and smile, clouds drift
Try this promptPlace these headphones on sunlit marble, orbit slowly, premium product shot
Try this promptThese prompts show the kinds of instruction Gemini Omni responds to. Run any of them in the Studio to see your own render.
Workflows
Every way to run Gemini Omni.
- Video editingConversational edits on an existing clipOpen workflow
- Text to videoPrompt to video, audio in the same passFrom 30 credits
- Image to videoAnimate a still, optional last frameFrom 30 credits
- Reference to videoCarry a subject or scene across shots from referencesOpen workflow
Google has not published a public rate for Gemini Omni, so cost is set by the workflow the Studio runs. Credits come from one shared balance across image, video and speech, and the exact estimate is shown in the Studio before you generate.
Built for
- Rapid post-production iteration without timeline software
- Background replacement that preserves the subject
- Wardrobe and outfit changes with identity locked
- Style transfer that holds consistency across a shot
- Turning sketches, maps and product shots into finished footage
FAQ
Gemini Omni, answered.
Limits, audio, licensing and price — the things worth knowing before you spend a credit.
Edit video by describing the change.
Open the Studio, pick Gemini Omni Flash, and see the credit estimate before you spend anything.