Person Swap lets you place a different model into an existing product photo — no reshooting, no manual compositing. Vision’s AI handles the replacement and blends the result into the background you choose. After this guide, you’ll know how to select an actor, pick an environment, upload your source image, and generate a swapped output ready for your campaign.
This workflow is a great fit when you want to show the same product on a different model or in a different setting without scheduling a new shoot.
Before you start
- You have a Vision account and belong to a team. If you haven’t set one up yet, see the Team Onboarding guide.
- Your team has at least one Actor saved. The actor generator guide shows how to make one.
- Your source photos (the photos whose model you want to replace) are saved locally. Vision accepts PNG, JPG, WEBP, AVIF and HEIC files up to 10 MB each.
- Your team has enough credits to run a generation. Check your balance on the dashboard, or see Credit Topup if you need to add more.
1. Create a Person Swap project
Person Swap is a project type of its own. From the Vision Dashboard, click + New Project. Give the project a Name, set Type to Person Swap and click Create.
Vision opens the Person Swap editor. It has an Actor picker, an Environment picker, an Extra prompt field and a Photos to swap panel. The Generate swaps button stays disabled until you pick an Actor and add a photo.
2. Choose an actor
In the Actor picker, type the name of the model you want to use in the search box. Vision searches your team’s saved actors as you type.
When the result appears in the dropdown, click it to select that actor. Vision loads a set of reference images (assets) for the selected actor — the first one is selected automatically. You can click a different thumbnail if you want to use a specific reference shot.
Tip: Actors are the reference models your team has added to Vision. If the person you’re looking for doesn’t appear in the search results, ask a team admin to add them first.
3. Choose an environment (optional)
The Environment picker works the same way as the actor picker. Type a scene name into the environment search box, then click the result that appears.
Once selected, Vision displays the environment’s reference assets below the search box. An environment tells Vision what background setting to use for the swapped image — for example, a studio backdrop, an outdoor scene, or a branded location.
Tip: Skipping the environment is fine if you only want to change the model and keep the original background. The Environment field is optional. Generate swaps enables without it once an Actor and a photo are in place.
4. Add an extra prompt (optional)
Use the Extra prompt field to describe any details you’d like Vision to incorporate — for example, wearing a red jacket or holding the product at waist height. Keep prompts short and descriptive; one or two phrases work best.
5. Upload your source image
In the Photos to swap panel, click the upload area and select your source photos, or drag and drop them onto the panel. When the files are attached, click Upload.
A thumbnail of each photo appears under the upload area within a few seconds.
Tip: You don’t need to remove the background from your source photo first.
6. Generate the swap
With an Actor and a photo in place, Generate swaps becomes active. The cost in credits shows above the button. Click Generate swaps.
Vision shows “Person swap scheduled.” and adds a row per photo to the Results panel. Each row reads “Waiting for output…” until its image is ready.
7. Review your output
Your swapped image appears in its row in the Results panel, next to the source photo. Click the thumbnail to view the full result.
Click Accept to keep an image, Reject to set it aside or Regenerate to run it again. Show rejected brings rejected images back into view, and Regenerate rejected reruns all of them at once.
Tip: Each generation spends credits from your team’s balance. You can review what was charged in your Credit History.
Next steps
You’ve successfully swapped the person in a product image — your output is saved to the project and ready to download or use in your next campaign. From here you might run additional generations with different actors or environments to build out a full set of variations, or explore the Background Removal feature to cut out any image precisely before compositing it elsewhere.
All brand names are property of their respective owners and used for demonstration purposes only.