Watch this tutorial on YouTube

YouTube

What Pet Video Does

VisionStory Pet Video turns pet reference photos and a written scene into a short generated performance. A single pet can carry the scene, or you can add a second reference and write an interaction between them.

The example in this tutorial uses an orange tabby chef and a golden retriever sidekick. Their red chef outfit and blue bandana give the model simple visual anchors while the prompt moves through setup, action, and reaction.

Before You Begin

Use photos you own or have permission to use.
Prepare one clear image for each pet you want to feature.
Keep faces and bodies visible, without heavy obstruction.
Decide where the final video will be posted before choosing its frame.
1 Find a Starting Point

Browse Explore or Start Your Own Scene

Open AI Tools → Pet Video. The Explore gallery is useful for understanding the kinds of actions, camera framing, and comic beats the tool can create. Hover a sample to reveal Recreate.

Recreate loads that example's reference, prompt, and settings into the composer. Treat it as a starting structure: replace the source media and rewrite the scene so the result belongs to your pets.

Current VisionStory Pet Video Explore gallery with Recreate visible on a hovered example
Hover an Explore example to preview its idea and reveal Recreate.
2 Add Pet References

Use One Clear Image per Pet

Add the first pet photo to the composer. To build a shared scene, click the + button and attach the second pet as another reference.

Choose images where each pet is easy to identify. A front-facing or three-quarter view usually gives the model more useful facial and body information than a distant, blurry, or heavily cropped shot.

Two clear pet photos attached as separate references in the current VisionStory Pet Video composer
Keep each pet in its own reference image, then add the next reference with the plus button.
Keep roles distinct. Different colors, accessories, breeds, or silhouettes make it easier to preserve who is doing what. Avoid giving both pets nearly identical outfits when the scene depends on separate roles.
3 Direct the Story

Write Every Pet's Action in Order

A multi-pet prompt works best when it reads like a short shot list. Name the shared location, identify each pet, then describe one action at a time in the order the viewer should see it.

1
Characters + settingSay who they are and where they are
2
Actions in orderMake it clear who does what first
3
Reaction + continuityEnd on a reaction and preserve identities
Current Pet Video prompt field showing two pet references and an ordered multi-pet story
Give each pet a named role, write the interaction chronologically, and finish with a clear reaction.

Keep the action achievable within a few seconds. A focused scene gives the model more room to establish the characters, complete the action, and land the reaction.

4 Set the Output

Choose Duration, Frame, and Review the Estimate

Open the output settings beneath the prompt. Pet Video currently supports clips from 4 to 15 seconds. Use a shorter duration for one simple action; give a setup-and-reaction scene a little more room.

Choose the frame that matches where the video will live. The current panel includes landscape, portrait, square, and intermediate aspect ratios. In this tutorial, a six-second 9:16 result keeps both pets prominent for vertical playback.

Current VisionStory Pet Video composer and output panel showing duration, aspect ratios, 720p, and live credit estimate
The panel keeps duration, frame, current resolution, and the live estimate together before generation.
Trust the live estimate. The displayed cost changes with the selected duration and may change as product pricing evolves. Review the number beside Generate rather than relying on an older example.
5 Generate and Review

Open the Finished Clip and Check the Story

After confirming the references, prompt, duration, frame, and current estimate, click Generate. The completed task appears in Recent Creations.

Open the task in Video Viewer and watch it from beginning to end. Check that the pets remain visually distinct, the assigned roles do not switch, and the action order matches the prompt.

Completed Pet Video task beside the real VisionStory Video Viewer playing the cat and dog result
Open the completed task in the real Video Viewer and review identity, action order, and the final reaction.

Quality Checklist

The pets look distinctColors, accessories, and body shapes remain recognizable throughout the clip.
The roles stay assignedThe prompt consistently names which pet performs each action.
The action fits the runtimeThe story contains a focused setup, action, and reaction instead of too many beats.
The frame suits the destinationImportant faces and actions remain visible in the chosen aspect ratio.

Frequently asked questions

  • Yes. Attach another pet image with the plus button and give every pet a distinct role in the prompt.