Different places. One place together.

A new portrait is different from a photo collage

A collage arranges existing photographs next to one another. A generated family portrait creates a new imagined scene using the references. Faces, poses, clothing and lighting can be reinterpreted, which means you need to check the result rather than treating it as a pixel-perfect merge.

Use a collage or conventional editing if keeping the exact original pixels is the priority. Use a portrait direction when you want to imagine everyone in a shared setting and are comfortable with the result being clearly AI-created.

How should I organize the reference photos?

Start with one person at a time. Pick their clearest image and put any supporting angle on that same person’s card. Do not combine several people into a single new reference collage: distinct participant cards make it easier to communicate who belongs where.

• Avoid tiny screenshots and photos where another person covers the face.
• Prefer similar, recent ages unless the intention is a specific remembered chapter.
• Use a full-body reference for a full-length composition when available.
• Check the roster before paying or generating.

What if the photos have different lighting?

Different rooms and cameras are expected. What matters most is that facial details remain visible. Strong coloured light, deep shadows, blur and filters hide information that the generator cannot reliably recover. A plain, clear source is often more useful than a dramatic photograph.

Choose a shared setting and clothing palette once the roster is ready. After generation, review one face at a time, then the group as a whole. If someone is too short or an outfit is wrong, describe that person using their participant label in the included revision. Keep everyone else unchanged when that is what you want.