Gemini video generator: how to make a video and edit it
Last updated 2026-10-20
The Gemini app makes video with Gemini Omni, which Google says will replace Veo 3.1 there. Go to gemini.google.com on a computer, choose Create video, enter a prompt, optionally add images or a video, and submit. Google lists 10-second videos with audio, and a Google AI plan is required. A finished video still needs editing.
How to make a video in the Gemini app
Google's Help article says Gemini Apps make video with Gemini Omni, and its Gemini app page says Omni will replace Veo 3.1 in the app. Google's overview names it Gemini Omni 1.1 Flash. Google's Help article gives these steps for a computer.
You can also refine the prompt in a normal chat first and generate from there.
- Open Gemini. Go to gemini.google.com on a computer. On Android, iPhone or iPad, open the site or the Gemini app, tap Menu, then Videos.
- Choose Create video. Open the sidebar and select Create video. You can pick a template for a preset style.
- Write the prompt. Add images or a video if you want them, and ask for audio if you want it. Text prompts default to landscape; uploaded media sets the ratio to match.
- Submit and wait. Generation can take a few minutes, and you cannot use that chat while it runs.
- Download. Under the video, choose Share, then Download video or Share on YouTube.
| Question | Google's answer |
|---|---|
| Model | Gemini Omni 1.1 Flash, which will replace Veo 3.1 in the app |
| Length | 10-second videos |
| Sound | Native audio; ask for it in the prompt |
| Inputs | Text, plus one video and up to five images |
| Who can use it | Personal accounts need a Google AI plan; work or school accounts need a qualifying Workspace license; not yet open to users under 18 |
| Marking | An imperceptible SynthID watermark |
Write a prompt for a 10-second video
Ten seconds is short, so a prompt should describe one moment. A useful order is subject, action, camera, then sound.
- Subject and action. Say who or what is in frame and what happens. One event per prompt keeps the clip readable.
- Camera. Google says prompts can change camera angles, so name the angle you want first and adjust in the next turn.
- Photos. You can add up to five. Google asks you to upload only photos you have the right to use.
- Your own likeness. Google's avatars let you use your own voice and likeness. Type @ followed by your Google username in the prompt.
Editing inside Gemini versus editing a video
Omni edits by conversation. Google says you can upload a video and ask for edits, remove or replace objects and characters, change camera angles or alter the scene, and keep going over several turns in one conversation. Google adds that you cannot upload a video for editing in the EEA, Switzerland, the UK or some US states.
Those edits change one clip. A finished video needs another kind of editing: put clips in order, trim them, add captions and music, and export. Because the app makes 10-second videos, even a 30-second video is at least three clips.
Other Google routes to the same models
Google's Omni page lists the Gemini app, Flow, YouTube Shorts, Google Vids, the Gemini API and Google AI Studio, where the model ID is gemini-omni-1.1-flash. Access needs a Google AI subscription, with features that differ by tier and geography. Omni content made or edited in the Gemini app, Flow or YouTube carries SynthID and C2PA Content Credentials.
Veo 3.1 stays listed on its own Google page for Flow, Vids, AI Studio and the Gemini API.
Use Google's video models in Woven
Woven lists two Google video families: Gemini Omni Flash 1.1, and Veo 3.1 with Veo 3.1 Fast. Omni Flash 1.1 shows text-to-video, image-to-video, reference-to-video and an edit mode. Veo 3.1 shows text, image, first-and-last-frame and reference modes. You ask in chat, and hosted generation is billed to Woven credits, with each model's rate on the pricing page.
Ask in chat
- Make a Gemini Omni Flash 1.1 clip of my cat noticing a drone, handheld camera.
- Generate the same shot with Gemini Omni Flash 1.1 and Veo 3.1 so I can compare the two.
- Use the edit mode of Gemini Omni Flash 1.1 on this generated clip and make the jacket red.
Or do it by hand
- 1Ask in chat. Name the model and describe the shot; the agent finds it on the live list.
- 2Pick the mode. Start from text, from one photo, from a reference, or ask for an edit of a clip you already generated.
- 3Import. Each result lands in the generated folder of your workspace. Drag it to the timeline, or use Add, then Video, and select it there.
- 4Build the video. Put the clips in order on the Picture row, trim by dragging their edges, and press ⌘B to split. Close gaps by dragging later clips left.
- 5Caption and export. Choose Add, then Generate captions, adjust each clip's Volume, and choose Export.
Common questions
Yes. Google's Gemini app page lists video generation powered by Gemini Omni, with 10-second videos and native audio.
Gemini Omni, named Gemini Omni 1.1 Flash on Google's overview page. Google says it will replace Veo 3.1 in the app.
Not on its own. Google's overview says you need a Google AI subscription, with features differing by tier and geography, and that college students can get a Pro plan free for one year.
Google's overview page says the app creates 10-second videos. For anything longer, join several clips in an editor.
Google says every video made in the Gemini app carries SynthID, an imperceptible watermark. The page does not mention a visible one.
Google's Help article says to open gemini.google.com or the Gemini app on Android, iPhone or iPad, tap Menu, then Videos, enter a prompt and tap Submit.
Sources
- Video generation in the Gemini app - Google, checked October 20, 2026
- Generate videos with Gemini Apps - Gemini Apps Help, checked October 20, 2026
- Gemini Omni - Google DeepMind, checked October 20, 2026
- Veo - Google DeepMind, checked October 20, 2026
Try Woven free for 3 days
Then $25/mo billed yearly ($299/yr) or $39/month. Native Mac app for short-form video.
Paste into Claude Code to install and connect Woven.