1. What happened: free generation and finer control over clips

On September 23, 2026, Google announced an expansion of AI video generation in Vids. Google or Google Workspace accounts can generate video with Gemini Omni 1.1 Flash for free. New controls include setting clip duration, extending a scene and generating 1080p footage.

Vids is a video-creation tool. These changes let you generate a scene and adjust its duration and content, turning an idea into a viewable first draft. Free generation has capacity limits; paid personal or Workspace plans offer more capacity. Check your account for the actual allowance.

2. How it helps: see the idea before deciding how to present it

Imagine preparing an opening video for a community reading group: a sunlit desk and a hand turning a page could set the mood. Instead of first filming, sourcing material and editing it, you could generate a draft to see whether the scene and pacing fit. This is a suggested scenario, not a production we tested.

Duration control can help match narration: if viewers need more time, you can try extending the scene rather than immediately cutting to another shot. Keep confirmed event names, venues and times separate from the imagined footage. A generated shot can establish atmosphere; it is not evidence of a real event or product.

3. How to use it: generate one scene and review the draft

The announcement’s starting point is Create AI videos in Google Vids on a computer. Try this suggested workflow:

  1. Sign in and check access. Open Vids on your computer, look for Create AI videos, and check your available generation allowance. If the entry point is absent, check account and organization conditions first.
  2. Describe one scene. For example: “A bright community reading room. A pair of hands slowly opens a book on a desk. Steady camera; no event dates or addresses.” This suggested description focuses on one visual idea.
  3. Set the duration and generate. Match the length to your opening’s pace. If you extend the scene, inspect the join and the consistency of people and objects.
  4. Review the draft, then add factual information. Play the whole clip and check movement, text and transitions. Use confirmed event details in the final content rather than asking the model to invent a venue or contact information.
  5. Watch the finished video before sharing. Check that captions, facts and images agree, and disclose the use of AI-generated footage.

Google says generated clips carry an imperceptible SynthID watermark. It helps identify AI-generated material; it does not show that a scene really happened. If you need more generation capacity, compare the allowances and prices shown in your account after trying a small task.