A multiple face swap changes more than one face in a single photo or video clip. In practice the phrase covers two different jobs. One is giving every person in a group shot the same new face, which is the classic "everyone is me" meme. The other is giving each person a different new face. They need different tools, so this guide explains both, shows how to do the first in aideepfake in about a minute, and says plainly when you need something else.
The two meanings of "multiple face swap"
Before you upload anything, decide which result you want:
- One face on everyone. You have one source face, for example your own selfie or a celebrity from a source library, and you want it on every person in the frame. Group photos, team pictures, wedding tables, a crowd scene from a movie clip.
- A different face for each person. You have a group photo and several source faces, and you want person A to become face 1, person B face 2, and so on. This is per-face mapping.
aideepfake does the first job in a single render, for photos and for videos. It does not map different source faces to different people in one render. Every face it replaces takes the identity of the one source face you uploaded. If you need per-person mapping, skip to the section on dedicated tools below. It lists several services and their limits as published on their own pages.
How to put one face on everyone in a group photo
You can do this with the standard photo face swap tool. You only need to change one setting.
- Upload the source face. Use a sharp, front-facing photo of one person, evenly lit, with nothing covering the eyes or mouth. If the source image contains several people, crop it to the face you want so there is no ambiguity.
- Upload the group photo as the target. JPG, PNG and WebP are accepted. Bigger faces swap better. If the group is a tiny cluster in the middle of a wide shot, crop the photo first.
- Open the advanced options and find "Face selection." Set Selector mode to many. By default the tool swaps only one face, the largest one. "Many" tells it to replace every face it detects.
- Generate and check every face. Zoom in on each person. Faces in profile, faces partly hidden behind another head, and very small faces at the back of the group are the ones most likely to look off or be skipped.
A photo render costs the same flat number of credits however many faces are in the frame. The current price per render is on the pricing page. That matters for big groups, because some services bill per face (examples below).
How to swap every face in a video
The video face swap tool works the same way, with one difference in its default. For video it picks the largest face in the first frame and follows that person through the clip, leaving everyone else untouched. That is what you want for a single-actor scene, but not for a group.
To swap the whole cast, set Selector mode to many in the advanced options before you generate. Every detected face in every frame is then replaced with your source face. MP4, MOV and WebM files are accepted. Longer clips are capped. The limit is shorter when face or frame enhancement is on, because enhancement processes every frame. The tool shows the cap for your settings before you submit.
Practical tips for multi-person video:
- Choose clips where faces stay visible. People turning fully away, motion blur and heavy shadows cause flicker on those frames.
- Keep occlusion masking on. It keeps hands, microphones and hair that cross in front of a face on top of the swap, so a wave or a sip of coffee does not punch a hole in the new face.
- Trim to the part you need. A shorter clip renders faster, and it keeps you under the cap when enhancement is on.
If you want famous faces rather than your own, the celebrity face swap library lets you pick a source face without uploading one. "Many" mode works the same way with a library face.
Swap only some of the faces
Sometimes "everyone" is too much. The Face selection group has two controls that narrow it down, and they work with "many" mode:
- Gender filter. Set it to female or male and only faces the detector classifies that way are replaced. Example: put your face on every man at a bachelor-party table and leave the women unchanged.
- Age range. "Age from" and "Age to" restrict the swap to faces the model estimates fall inside that range.
Both filters depend on the model's estimate, which can be wrong on stylised photos, heavy make-up or unusual lighting. Check the result before you share it.
If you want exactly one specific person rather than a group, set Selector mode to reference. Pick a Selector order (for example left to right). Then set Reference face position to that person's place in the order, counting from 0. This swaps the chosen person and leaves everyone else as they were.
When you need a different face for each person
If your goal is per-face mapping, use a tool built for it. On 2026-11-03 we checked several vendors' own pages. Here is what each one says it offers:
| Service | What its page says (checked 2026-11-03) |
|---|---|
| Facecam (facecam.ai) | Up to 6 faces per swap in photos, GIFs and videos. Each face swapped counts as one swap. |
| Pincel (pincel.app) | Up to 10 faces per photo. The page warns quality may drop when faces are small. |
| faceai.art | Up to 10 faces per photo in one pass. Free previews carry a watermark. |
| Facy (facy.ai) | Detects up to 20 faces per photo. Requires a free account. |
| Magic Hour (magichour.ai) | The free photo tool supports up to three replacement faces. More requires the dashboard. |
| Vidwud (vidwud.com) | Its pricing page bills 7 to 10 faces in a photo at 2 credits per image, and 4 to 10 faces in a video at "faces × 2" credits per 15 seconds. |
These limits and prices change often, so check each vendor's page before you pay. Two patterns are worth knowing. First, per-face billing gets expensive for big groups. Second, most per-face tools cap the number of faces, so a large group may still need two passes. For a wider look at the market, see our side-by-side face swap tool comparison and our list of the best video face swap tools.
Getting a clean result with many faces
Group shots are harder than single portraits, for predictable reasons:
- Face size. The detector works on a downscaled copy of the frame, so a face that is only a few dozen pixels wide may be missed or come out soft. If the back row is skipped, crop the photo so the faces are larger, then swap again.
- Detector score. In advanced options, Detector score sets how confident the model must be before it counts something as a face. Lowering it a little can pick up small or angled faces. Lowering it too far can make it "find" faces in patterns or posters.
- Angles. Swaps hold up best on faces within roughly a three-quarter turn. A full profile has too little visible face to map well.
- Lighting mismatch. One source face is being blended into several lighting conditions at once. A source photo with soft, even light adapts best.
- Enhancement. The face enhancer sharpens swapped faces, which helps when they are small. On video it shortens the maximum clip length.
Consent and the law still apply to every face
A group photo contains several people, and each of them has a stake in how their image is used. Swapping a face onto a friend's birthday photo as a joke inside a group chat is very different from posting an edited image of a real person in a way that could deceive or humiliate them.
Two points to keep in mind:
- In the United States, the TAKE IT DOWN Act (Public Law 119-12, signed on May 19, 2025) makes it a federal crime to knowingly publish non-consensual intimate images, including AI-generated "digital forgeries." Platforms covered by the law must remove reported images within 48 hours of a valid request, and the FTC began enforcing that duty on May 19, 2026. Sources: Congress.gov, S.146 and the FTC announcement.
- App stores and payment providers are tightening their rules too. In November 2026 WIRED reported on hundreds of ordinary-looking apps that hid abusive deepfake features. Several were removed from app stores (WIRED, 2026-11-02).
Good practice is simple. Get permission from the people whose faces you use or change. Do not make sexual or degrading content of real people. Label edited media when you share it publicly. If you are unsure whether a clip you received is real, our guide on how to detect deepfakes covers the tell-tale signs. What is a deepfake? explains the underlying technology.
Multiple face swap for developers
If you are building a group-photo feature into your own app, the face swap API accepts the same advanced options as the web tool, including the selector mode. One request with "many" replaces every detected face with the source face you send. The API is priced per render, like the website.
FAQ
Can I swap multiple faces in one photo for free? You can try aideepfake's photo face swap with free credits, and free results carry an aideepfake.io watermark. Any purchase unlocks the switch that removes it. See the free face swap page for what the free tier includes.
Does aideepfake let me choose a different face for each person? Not in one render. Every face replaced in a render takes the single source face you uploaded. You can target one specific person with the reference selector, or use a per-face tool from the table above.
How many faces can be swapped at once? aideepfake does not set a fixed face count for "many" mode. It replaces every face the detector finds. In practice the limit is how large and clear the faces are in the image.
Does it work on GIFs? No. Targets must be photos (JPG, PNG, WebP) or videos (MP4, MOV, WebM).
Is the result watermarked? Free results carry an aideepfake.io watermark. Paid users can switch it off.
The short version
If you want one face on everyone, open the AI face swap tool. Upload a clear source face and your group photo or clip, set Selector mode to many, and generate. Use the gender or age filters to swap only some people, or reference mode to swap just one. If you need a different face for each person, pick a per-face tool and check its face limit and per-face pricing first.