To add a person in photos with AI, upload the photo you want to keep and a clear reference image of the person you want to include. Use an AI photo editor that supports reference images, specify where the person should appear, and ask it to preserve everyone already in the picture. If the editor offers a selection or mask, use it to guide the addition into an open space.
A family member missing from a festive portrait or the friend who took the group photo can become part of a new composite. The challenge is making them recognisable and physically believable without changing the rest of the group. A good result needs the right reference, sensible placement and a careful comparison with the original.
Choose an AI editor that accepts the right inputs
For a specific person, you need an editor that can use their photo as a reference. A text-only prompt such as “add a woman in a blue kurta” describes an appearance; it does not identify your sister.
Look for two capabilities: a way to supply both the base photo and the person's reference, and a way to direct placement. Some AI image editors let you select an area and use generative fill. Others combine multiple reference images through written instructions or a sketch. Adobe's reference-image workflow, for example, includes an option to place content into a selected area while preserving the background. That guides the edit, although you still need to inspect the generated result.
Before uploading a large family portrait, confirm that the tool supports adding a person rather than only replacing a face or creating an entirely new scene. These tasks produce different kinds of results.
Prepare the base photo and reference image
Choose a base image with space beside the group, an empty chair or a clear foreground area. Adding someone at an edge is often a simpler composition than squeezing them between people whose arms already overlap.
For the reference image, use a sharp photo with a visible face and natural lighting. A full-body image is useful if you want the added person's outfit and posture to remain recognisable. A close-up selfie provides little evidence about their clothing, height or legs, so the AI must invent those details.
Try to match the viewing angle. A front-facing reference is easier to work with in a group facing the camera than in a scene where everyone is photographed from the side. Similar lighting also reduces the amount of reconstruction required.
| Detail | Better starting point | Why it helps |
|---|---|---|
| Face | Clear, unobstructed features | Gives the model useful identity information |
| Body | A pose compatible with the group | Reduces awkward limbs and posture |
| Placement | Enough room for the whole person | Avoids crowding or replacing existing people |
| Light | Similar direction and softness | Helps the addition blend into the scene |
| Image quality | Original files where available | Retains more detail than compressed screenshots |
How to add someone to an existing photo
1. Assign each input a clear role
Upload the original group or scene as the base image. Add the person's photo as the identity reference using the editor's available controls. Where references have names or numbers, identify them explicitly in your prompt: “Image 1 is the base photo; Image 2 supplies the person to add.”
Avoid asking the AI image generator to blend both pictures equally. That can create a new background or a different arrangement instead of editing the photograph you wanted to keep.
2. Define the person's position
Choose a location with enough space for a believable body and contact with the ground or chair. In an editor with masking, select the intended placement area, allowing room for the body and a natural shadow. Keep the selection clear of existing faces where possible.
Without a mask, describe the position precisely: “standing at the far right of the back row” is clearer than “next to the family.” If supported, a placement sketch can make the instruction more concrete.
3. Use a prompt that protects the original photo
Start with this adaptable AI photo editing instruction:
Use Image 1 as the base photograph. Add the consenting adult from Image 2 at the far right of the group, standing on the same ground plane as the adjacent adult. Preserve their recognisable facial features, age, skin tone and outfit from Image 2. Match the scene's perspective, light direction, sharpness and shadows. Retain all original people, their expressions, clothing and positions, and the existing background.
Adapt the image labels to your editor. Preservation instructions guide generation; they do not guarantee that every original detail remains identical.
4. Inspect identity before approving the scene
Compare the added face with the reference at a useful viewing size. Check the eyes, nose, mouth, face shape and age. A polished portrait can still show identity drift.
Then compare every original person against the base photo. Reject a version that improves the addition but changes another person's expression, merges faces or removes someone from the group. For an important family photo, those changes matter more than a visually smooth background.
5. Correct one problem at a time
If the person looks too large, specify their position relative to someone at the same depth. If they appear to float, ask for grounded feet and a contact shadow matching the scene. If their face is unfamiliar, return to the original inputs and try a clearer reference.
A local correction in an AI editor can be useful when the face is right but a small detail is warped. Repeatedly regenerating the whole group increases the chance of changing people who already looked correct.
Prompts for three common photo situations
A family celebration
Add the adult from the reference photo to the empty chair at the left end of this family portrait. Preserve their identity and outfit. Match the seated eye level, indoor lighting and chair perspective. Keep all existing family members, decorations and the framing unchanged.
For a Diwali or wedding-related portrait, inspect embroidered clothing, jewellery and decorations closely. Fine patterns can turn into invented detail around the edited area.
A graduation group
Insert the person from the reference image at the right end of the front row. Keep their face recognisable and their gown consistent with the supplied reference. Match the scale of the adjacent graduate at the same distance from the camera. Preserve all original faces and readable details elsewhere.
Provide the actual graduation outfit as a reference if it matters. Asking the AI to guess a gown or stole can introduce incorrect colours and symbols.
A travel photo
Add the person from Image 2 standing beside the railing on the left of Image 1. Preserve the landmark and the existing people. Match the camera height, afternoon sunlight and shadow direction. Keep the added person behind the railing where their body overlaps it.
That final instruction addresses depth: a railing should naturally cover the parts of a person standing behind it.
Fix the details that make an addition look artificial
| Problem | What to check or change |
|---|---|
| The face resembles someone else | Use a clearer identity reference; reject identity drift |
| The person looks pasted on | Match light, colour, sharpness and shadow softness |
| Feet float above the ground | Specify placement and contact shadow on the correct surface |
| Hands or clothing merge | Simplify the pose and reduce overlap with other people |
| The group rearranges itself | Narrow the edit area or use a more local editing workflow |
| The scene's scale feels wrong | Compare people at the same depth and check perspective |
Where Jokepop fits
Jokepop offers playful photo effects and photo-based dance videos for social sharing. Explore its effects when your goal is a fun new creation from a photo. For inserting a named person into a specific existing group picture, choose a workflow with explicit reference-image and placement controls; a creative effect alone does not establish that capability.
Frequently asked questions
Can I add someone using only a description?
You can describe a new person, but a description alone cannot reliably reproduce a specific individual's identity. Supply their reference photo when the result needs to look like them.
Can AI keep the original group completely unchanged?
It can be instructed to preserve the group, especially with a local edit, but generated outputs may still change nearby details. Compare all faces and important objects with the original before keeping the result.
Can I do this on my phone?
Yes, with a mobile app or browser-based AI photo editor that supports the required inputs. Confirm that you can supply a reference image and guide placement before committing to the workflow.
What should I consider before sharing the composite?
Get permission from the person whose image you use and make clear that the photo is AI-edited when viewers could mistake it for a record of attendance. Review the service's upload and retention terms before submitting private family photos.
Start with one person, a clear reference and a believable space in the base photo. Once the face, scale and lighting work together, check the original group again. The strongest edit makes room for the added person while preserving the people and scene that made the photograph worth keeping.
