Quick answer: Google Vids now lets anyone with a Google or Google Workspace account generate AI video at no cost on desktop. Google says Gemini Omni 1.1 Flash can create new scenes in 1080p, extend scenes, preserve visual context across extensions, and let you set specific clip durations. Start at vids.new and choose Create AI videos.[1]
Fast setup checklist
- Use a desktop browser and sign in to a Google account.
- Open vids.new.
- Select Create AI videos.
- Describe one scene at a time using the prompt template below.
- Set the duration, generate the scene, then extend or revise it.
- Review every frame, claim, logo, face and caption before publishing.
What is free in Google Vids?
Google’s September 23 announcement says people with either a personal Google account or a Google Workspace account can use Gemini Omni 1.1 Flash in Google Vids at no cost. The free experience includes AI scene generation, exact-duration controls, scene extension, 1080p generation and upscaling of existing AI clips.[1]
| Feature | What it does | Practical use |
|---|---|---|
| 1080p generation | Creates new AI video scenes in full HD | Product teasers, explainers and social clips |
| Exact duration | Lets you specify the clip length | Fit scenes to a script, voiceover or ad slot |
| Scene extension | Continues a scene while preserving context | Build a longer sequence without restarting |
| Upscaling | Raises existing AI clips to 1080p | Make mixed timeline assets more consistent |
| Templates | Provides ready-made starting layouts | Launch videos faster without editing from zero |
Google also says personal Google AI plans provide more access, while Workspace Business and Enterprise plans include expanded generation pools and centralized administration. The announcement does not state one universal numerical free-generation quota, so do not assume “free” means unlimited.[1]
How to make your first free AI video
1. Open Google Vids
On a desktop computer, visit vids.new and sign in. Choose Create AI videos from the creation menu. Google’s launch instructions specifically point to desktop access.[1]
2. Plan a short sequence
Do not ask for an entire campaign video in one vague prompt. Break the idea into three to six scenes. Give every scene one job: establish the setting, show the product, demonstrate the action, present the result, or close with a call to action.
3. Use this copy-paste prompt template
Create a [DURATION]-second 1080p video scene. SUBJECT: [person, product, place or object] ACTION: [one clear action] SETTING: [location, time of day, background] CAMERA: [wide/medium/close-up], [static/pan/dolly/handheld] LIGHTING: [soft daylight/studio/neon/golden hour] STYLE: [realistic/product demo/editorial/animated] COLORS: [brand or mood colors] CONTINUITY: Keep [subject, clothing, product, lighting] consistent. AVOID: text artifacts, extra fingers, warped logos, abrupt cuts. OUTPUT: No on-screen text unless specified.
4. Generate, inspect and revise
Check the subject, hands, faces, product shape, branding, reflections and background movement. If the scene is nearly right, make one targeted change rather than rewriting every detail. This makes it easier to identify which instruction improved or damaged the result.
5. Extend only after the first scene works
Omni 1.1 can extend a scene while maintaining elements such as visual context, lighting, character appearance and environment. Use extension after you have a clean starting shot; otherwise, errors can continue into the next segment.[1]
6. Add captions and publish carefully
Proofread captions, verify factual claims and get permission for identifiable people, voices, trademarks and customer material. AI generation can speed production, but it does not transfer responsibility for accuracy, consent or usage rights.
Three ready-to-use prompt examples
Local business product clip
Create an 8-second 1080p realistic product video. A barista places a freshly prepared iced coffee on a clean wooden counter in a bright neighborhood cafe. Medium close-up, slow forward camera movement, soft morning daylight, warm brown and cream palette. Keep the cup shape and label position consistent. No on-screen text and no extra objects appearing.
App explainer background
Create a 6-second 1080p editorial technology scene. A small-business owner reviews a simple analytics dashboard on a laptop in a modern office. Over-the-shoulder camera, gentle side movement, natural daylight, blue and neutral palette. Screen content should remain abstract and unreadable so real interface graphics can be added later.
Community event opener
Create a 7-second 1080p welcoming community-event opener. Diverse adults arrive at a bright public library meeting room and greet each other. Wide shot moving slowly to medium shot, realistic documentary style, soft afternoon light, calm movement. Leave clean negative space on the upper left for a title to be added during editing.
A practical production workflow
- Write the message first: define the audience, one takeaway and one action.
- Draft a six-line script: hook, problem, demonstration, benefit, proof and next step.
- Convert each line into a scene: keep visual instructions specific and bounded.
- Generate low-risk visuals: avoid unsupported claims, fake interfaces and identifiable people without permission.
- Review continuity: compare product shape, wardrobe, colors and lighting between scenes.
- Add real brand assets later: insert approved logos, prices, screenshots and legal copy during editing rather than asking the model to invent them.
- Export and watch the complete video: inspect it once with sound and once muted with captions.
AI watermarking and disclosure
Google says every Omni 1.1 clip generated in Vids includes an imperceptible SynthID watermark in the video frames.[1] Google DeepMind describes SynthID as a watermark embedded directly into AI-generated content and says its video watermark is designed to survive common changes such as cropping, filters, frame-rate changes and lossy compression.[2]
An invisible watermark is not a substitute for a clear disclosure when context could mislead viewers. Add a visible “AI-generated” or “AI-assisted” note for realistic synthetic people, events, demonstrations or scenes that viewers might mistake for documentary footage.
Important limits and cautions
- Free does not mean unlimited: Google says paid plans offer more access but does not provide one universal free quota in the launch announcement.[1]
- Desktop is the documented starting point: the launch directions send users to vids.new on desktop.[1]
- Text-to-speech is not part of the current promise: Google says Gemini 3.8 Flash-Lite voiceover across more than 100 languages is coming soon. Do not treat it as universally available yet.[1]
- Generated facts are not evidence: manually verify prices, specifications, medical statements, legal claims and news details before publishing.
- Use authorized assets: obtain permission for customer media, employee likenesses, protected music and third-party branding.
Frequently asked questions
Is Google Vids free now?
Yes. Google says anyone with a Google or Google Workspace account can generate videos with Gemini Omni 1.1 in Google Vids at no cost. Paid Google AI and Workspace plans provide more access or larger managed generation pools.[1]
Can Google Vids generate 1080p video?
Yes. Google says Omni 1.1 in Vids can generate new scenes in full 1080p HD and upscale existing AI clips to 1080p.[1]
Do I need a Google Workspace subscription?
No. The announced free access includes personal Google accounts as well as Google Workspace accounts.[1]
Where do I start?
Open vids.new on desktop, sign in and select Create AI videos.[1]
Are Google Vids AI clips watermarked?
Yes. Google says every Omni 1.1-generated clip in Vids has an imperceptible SynthID watermark embedded in its frames.[1]
Is AI voiceover available?
Google announced that text-to-speech voiceover in more than 100 languages is coming soon. Availability may therefore differ when you open Vids.[1]
Last checked: September 24, 2026.
Official sources