Canva AI’s Video Clip feature makes a brief scene from a written prompt. Canva’s current help page says a generated clip can be eight seconds with audio or six seconds without audio. The prompt must be in English, even if the rest of your project uses another language. Treat the output as one shot for a longer edit, not as a complete narrated video.
Write a prompt that describes a shot
A useful prompt specifies the subject, action, setting, composition, camera movement, light, and mood. Keep one main action in the clip. For example:
> Vertical close-up of a ceramic cup on a wooden table as steam rises in morning window light; slow camera push-in, quiet natural colors, no text or logos.
This gives the model visual direction without asking it to fit a whole story into a few seconds. Avoid asking the model to render exact captions, prices, or product labels. Text inside generated imagery can be misspelled or distorted; add important copy in the editor after generation.
Generate and edit the clip
- Open Canva AI from the Canva home page.
- Select Video Clip.
- Enter an English prompt that describes one short scene.
- Optionally attach a reference image if you have the right to use it and Canva offers the option.
- Review the generated clip for unwanted objects, visual glitches, and whether the action matches the prompt.
- Download the clip or choose Use in Canva Editor to trim it, add text, and place it in a larger video.
- Add your own narration or licensed music only after checking how the clip’s generated audio behaves. Canva says audio in clips with sound cannot currently be manually changed in the generation interface.
The interface and available settings may change. If you do not see Video Clip, check the current account features and region rather than following an old screenshot. Canva’s AI usage guidance says Free users have a limited shared allowance for some tools; usage is not unlimited, and the allowance details depend on the feature and plan. The app may ask you to upgrade when a feature or allowance is not available on your account.
Build a complete short video from several clips
For a 20- to 30-second explainer, generate a few different shots only if each contributes new information. Combine them with original footage, title cards, narration, and a clear beginning and end. Keep visual continuity by repeating a small set of details—color palette, lens style, or setting—in each prompt. Then inspect cuts for mismatched lighting or changes in the subject’s appearance.
If you already have photos or footage and mainly need help arranging them, Canva Magic Video is a better fit than text-to-video. This Magic Video workflow describes that option. For a free Google account route, see Google Vids and Gemini Omni.
Before publishing, verify that your subject is not being represented as a real person or event in a misleading way. Do not generate a recognizable person’s face or voice for an endorsement without permission. Keep a copy of the source prompt and the edited final file, and disclose synthetic elements when required by the platform or context. The fact that a clip is generated with AI does not make its music, reference images, or resulting post automatically rights-cleared.