Gemini Omni 1.1 Flash: Video Edit
Edit videos through natural language conversation with Gemini Omni 1.1 Flash. Upload a clip (10 seconds or less) and describe the change: swap backgrounds, restyle scenes, add or remove elements, and the model applies the edit while preserving everything you did not mention. Add \"Keep everything else the same\" for maximum consistency; simple prompts work best. Outputs video with native audio at 16:9 or 9:16, with 1080p/4K output available. Ideal for remixes, style changes, and iterative refinement without re-generating from scratch.
简介
This ComfyUI workflow lets you edit a short video by simply describing the change you want. LoadVideo ingests your clip (the template includes input/hooded_wanderer_planet_bridge.mp4 as a ready-to-run test), and GeminiVideoOmniV2 sends both the video and your plain‑language instruction to Gemini Omni 1.1 Flash. The model applies targeted edits—like swapping backgrounds, restyling a scene, or adding/removing elements—while leaving everything else untouched.
你需要提供
- 正在读取工作流
在 ComfyUI 中运行。 此工作流无法在本页面运行。你可以在 Comfy Cloud 中打开（已预先加载所需内容），或下载后在自己的 ComfyUI 中使用。
在 Cloud 中试用常见问答
Keep inputs to 10 seconds or less. Shorter clips process faster and maintain better temporal consistency. Longer inputs may be slower or less stable depending on API limits.
Yes. The workflow is designed to export the edited video with its native audio via SaveVideo. If your source has no audio, the output will be silent.
Use simple, specific instructions and include “Keep everything else the same.” Avoid bundling many changes at once, minimize rapid camera motion, and keep the subject’s key attributes explicit (e.g., clothing, colors) if they must not change.
Yes. Set the aspect ratio to 9:16 in GeminiVideoOmniV2 (and select 1080p or 4K as needed). The workflow will render a vertical output with the edit applied.