Gemini Omni 1.1 Video Edit (Video-to-Video Editing)
What is Gemini Omni 1.1 Video Edit?
Gemini Omni 1.1 Video Edit is Google's instruction-based video editing model, exposed on Segmind as a simple video-to-video API. You upload a clip of ten seconds or less and describe a single change in plain English, and the model applies that edit while preserving everything you did not mention. It is the editing mode of Gemini Omni 1.1 Flash, Google's production-ready multimodal video model that reasons about physics, lighting, and real-world context to keep edits coherent from frame to frame.
Because the model understands the whole scene, changes like a full restyle or an object swap stay consistent across the clip instead of flickering between frames.
Key Features
- •Text-driven editing: restyle footage, add or remove objects, change lighting, or rewrite on-screen text from one instruction.
- •Subject and scene preservation: unmentioned elements, motion, and composition stay locked.
- •Resolution control: 360p draft for fast previews, 720p default, and 1080p or 4k upscaled finals.
- •Native audio: outputs return a synchronized audio track.
- •Reproducibility: fix a seed to repeat a result, or leave it random for variation.
- •SynthID watermarking on every output for AI provenance.
Best Use Cases
Gemini Omni 1.1 Video Edit is built for fast, iterative creative work. Use it to restyle live footage into anime, Pixar-style 3D, or claymation looks; swap or recolor products for ad variants; change backgrounds with lighting that reacts naturally; add wardrobe or props to a subject; or clean up unwanted objects. In testing, a photorealistic clip restyled to Japanese animation kept the subject, pose, framing, and camera motion intact with no artifacts, making it well suited to social clips, mood boards, and marketing concepts where you refine one idea at a time.
Prompt Tips and Output Quality
Keep prompts simple. Describe only the change and add "Keep everything else the same" to protect the rest of the shot, because overly detailed prompts cause unintended edits. Put negatives directly in the prompt, for example no dialogue or no extra people. Preview ideas at 360p, then rerun the winner at a higher resolution. Input clips must be ten seconds or less, and longer requests are clamped to ten seconds of output.
FAQs
What inputs does Gemini Omni 1.1 Video Edit need? A video URL of ten seconds or less and a text edit instruction; resolution and seed are optional.
Does it keep the rest of the video unchanged? Yes. Name only the change you want and add "Keep everything else the same" to preserve the scene.
What resolutions are supported? 360p, 720p, 1080p, and 4k, where 1080p and 4k are upscaled from the 720p render.
Can it edit voices or add dialogue? No. Voice editing and audio references are not supported; describe the audio you want in the prompt instead.
What languages work best? English is fully supported; other languages may work but have not been evaluated.