Create and Edit Short Videos Using Gemini Omni Flash Conversational AI
Edit 720p videos conversationally using the new Omni Flash model.
Leverage gemini-omni-flash-preview via the Interactions API to generate 3–10 second videos and refine them using natural language commands.
The Scenario
A social media manager needs to create a short, high-quality promotional video and wants to tweak specific details without starting the production from zero.
Before & after
Creating and iterating on short videos traditionally required complex video editing software and manual keyframing, taking 2–4 hours for a 10-second clip.
With Gemini Omni Flash and the Interactions API, you can generate a 720p video from text and refine it through chat in about 2–3 minutes.
The Prompt
Generate a 5-second 720p video of [SCENE_DESCRIPTION]. Once generated, I want to edit it by [CONVERSATIONAL_EDIT_INSTRUCTION, e.g., 'adding more sunlight to the background'].
Gemini Omni Flash (preview) allows for conversational video editing, meaning you can describe changes to a video rather than re-rendering from scratch.
Source
Release notes | Gemini API | Google AI for Developers"generate 3–10 second videos at 720p from text descriptions... and then conversationally edit and refine the outputs."