To swap a face in another photo, you need two different inputs: the target picture containing the scene you want to edit, and the portrait reference showing the face you want to use. In faceswap, upload the target first, choose your custom reference, then generate and inspect the result.
Open the online photo face swap tool when your images are ready. This guide follows the current still-image workflow; for moving footage, use the separate music video tutorial.
Photo face swap or video motion transfer?
An AI photo face swap starts with two still images. A video project needs an MP4 and portraits assigned to its scenes. Choose the photo tool for a profile picture, a single stage photo, or one edited frame; use the video workspace when movement and source audio matter.
The video trial and photo credits have different rules. Read the AI video face swap trial and cost guide before assuming that a free video trial also covers additional photo generations.
Choose which photo plays which role
Mixing up the target and reference is an easy way to produce an unexpected composition. Decide which image supplies the scene before uploading either one.
| Input | What it supplies | Example |
|---|---|---|
| Target photo | The picture and composition to edit | A single-person stage photograph |
| Face reference | The identity cues for the edit | Your clear portrait |
| Result | A generated version of the target | An AI-edited stage image using your reference |
This example describes the input roles, not a guarantee that every item in the target will remain unchanged. Facial expression, hair boundaries, lighting, and nearby details can vary in generated results.
The photo editor accepts JPG, PNG, and WebP files up to 8 MB per image. Start with a clearly visible face in the target and a reference with a compatible angle. If you are comparing several portraits, use the photo quality scorecard before submitting.
Step 1: Upload the picture you want to edit
In the photo workspace, select the target picture. Check the upload preview: it should show the scene you want in the finished image, not just the portrait intended as your reference.
For a group photo, crop a copy around the intended person first. The current tool does not offer a precise selector for each person in the group. Keep the original file separately so you can try a different crop without repeatedly compressing it.
Avoid a target where the face is tiny, hidden behind a microphone, or heavily blurred. A clear full-resolution original is more useful than a screenshot of the same scene.
Step 2: Select your own face reference
The tool starts with a built-in Charlie Kirk reference. To put your own face on another picture, use the custom face option to upload your portrait. Confirm that the workspace says your custom reference is active before starting.
A practical first choice is a close portrait with visible eyes, nose, and mouth. Try to match the target head angle: a front-facing reference may provide less useful detail for a strong profile. If the target is smiling, a reference with a compatible expression is worth trying.
For the built-in character workflow, the Charlie Kirk Face Swap tutorial explains that specific use case. You do not need to use that reference for a custom photo swap.
Step 3: Sign in and check the credit cost
Each photo generation currently costs 4 credits. New accounts receive 4 starter credits after sign-in, valid for one year. Check the displayed balance and cost before submitting; do not assume the video trial applies to photo editing.
Submit once, then follow the progress indicator. Processing time varies with the input and provider queue. If you encounter an upload error, check the file format and size before trying again.
Step 4: Review before downloading
Inspect the result at normal viewing size, then zoom in on these areas:
- Eyes and mouth: check direction, shape, teeth, and expression.
- Face boundary: look for a hard seam near the jaw, forehead, or ears.
- Lighting and skin: compare the edited face with the neck and surrounding scene.
- Nearby objects: inspect glasses, hair, hands, text, and background lines.
Change one input when testing an improvement. For example, keep the target and choose a better-matched reference. If several areas are distorted, the artifact troubleshooting guide can help you decide what to change first.
Download after reviewing the image. Use photos you have permission to edit, and identify the result as AI-edited when sharing it in a context where viewers could mistake it for an original photograph.
Common questions
Can I swap faces using two photos?
Yes. One is the target picture; the other is the custom portrait reference. This produces one edited image. It does not simultaneously exchange both faces and return two separate pictures.
Do I have to install an app?
The photo workflow runs through the website. You upload the files in your browser and sign in to generate and download the result.
Why does the result still resemble the built-in face?
First check which reference was active when you submitted. Uploading the target photo alone leaves the built-in reference selected. Choose your custom portrait before generating another image.
Can I use a video in this photo editor?
Use the separate video motion transfer tool for MP4 footage. It uses scene-by-scene portraits and may change more of the character and scene than a still-photo face swap.
Ready to make your first version? Return to the photo face swap workspace, confirm the target and reference, and review the cost before generating.
