Gemini Omni is now available directly inside Google Vids, Google's video creation tool. Beyond generating clips from a prompt, it can rework footage you already have based on a written instruction. Personal avatars, which let you cast a digital version of yourself, arrived at the same time. The staged rollout began on July 16, and the audience widens from August 5.

Omni Flash raises the quality of generated clips

The first change is the quality of the clips you can build inside Vids. Generation runs on Omni Flash, Google's latest model, which the company describes as delivering improved text rendering, more convincing physics, and greater overall realism than previous models.

On-screen text and the behavior of objects in motion have long been the weak points of video generation. When either breaks down, even a short clip reads as synthetic at a glance. Vids is used heavily for slotting video into business decks and internal material, so improvements in those two areas are easy to notice in practice.

Editing by typing what you want changed

The second pillar is text-driven editing. Because Omni understands what is happening in the footage, you can make changes without working through a timeline in detail.

Google's examples are concrete: fix the color grading, restyle the whole clip in anime, or strip out a New York siren that leaked into the background — each from a single written instruction. That removes a round trip of reshooting material or exporting it to a separate editor.

There is a regional catch. Editing videos that were not AI-generated with Omni is unavailable at launch in the European Economic Area, Switzerland, the United Kingdom, and the US states of Texas and Illinois. Those are jurisdictions with tighter rules around generative AI, likeness, and data handling, so Google has drawn a cautious line around live-action footage.

Personal avatars built from a face and voice recording

Personal avatars landed alongside the Omni update. You start in Vids on a computer, pick Personal avatar, scan the QR code with a phone or tablet, and follow the prompts to record your face and voice. Once that verification is done, Vids builds a digital presenter that reproduces your appearance and voice. From there you describe how the avatar should move and speak, add a script if you need one, and it delivers the lines on your behalf.

For videos where only the content changes — internal announcements, product walkthroughs — the filming step disappears entirely. On the other hand, an avatar becomes an asset tied to a person's face and voice, so organizations need to settle who may create one and who may reuse it.

Conditions apply here too. At launch, personal avatars support English only and are limited to users aged 18 and over. They are not offered in the European Economic Area, Switzerland, or the United Kingdom.

Rollout schedule and eligible editions

Availability arrives in stages. Rapid Release domains started on July 16, Scheduled Release domains start on August 5, and in both cases the feature can take up to 15 days to appear. If it has not shown up in your environment yet, waiting is a reasonable option.

Eligible editions include Business Starter, Standard, and Plus; Enterprise Starter, Standard, and Plus; Education Plus; Enterprise Essentials and Enterprise Essentials Plus; and the Nonprofits edition. On the consumer side, Google AI Pro and Ultra subscribers are covered. Add-ons in scope are Google AI Pro for Education, Teaching and Learning, and AI Expanded Access, and users with AI Expanded Access get higher usage limits for Omni in Vids.

Omni generation and editing have no admin toggle: once the edition requirements are met, they become usable as soon as the rollout reaches the domain. Personal avatars work differently — they are on by default, but admins can turn them off or on at the domain level from the Admin console.

Summary

Bringing Gemini Omni into Google Vids improves both the creation side and the correction side. Omni Flash sharpens text rendering and physical motion, while written instructions now cover color grading, anime restyling, and background noise removal. Personal avatars offer a way to skip filming altogether, but both live-action editing and avatars carry regional and language limits, so what you can actually use depends on your environment. Rapid Release domains started on July 16 and Scheduled Release domains start on August 5, with up to 15 days for the feature to appear.

References: Google Workspace Updates: Omni in Vids / Google Workspace Updates: Personal avatars with Gemini Omni in Vids / Google Docs Editors Help: Create, use & manage your personal avatar