AI Family Photo Time Jump

One group or childhood photo in JPG or PNG. Faces should be sharp, evenly lit and roughly the same size in the frame. Heavy shadow, motion blur, sunglasses or faces turned far from the camera all cost identity accuracy.

The people-count control goes from one to eight, and naming the real number is what keeps every face in place. Two to five is where it works best; busier group shots drift. For a single face, the AI Aging Filter at /image/aging-filter is the focused tool.

No. This is entertainment, not a forensic or medical prediction. Real aging depends on genetics, sun exposure and lifestyle, none of which a photo can reveal. The result applies documented average aging markers — treat it as a fun picture, not a forecast.

That is the hard part of this edit, which is why GPT-Image 2.5 comes in two tiers here. Flare is fast and free; Sunburst is the precise tier built for holding several identities at once, and is Pro. Both keep the same people, arrangement and framing.

Rephotography means re-shooting a scene from the same viewpoint years later. Pick the matched-framing treatment and the edit holds the original pose, crop and lighting, so the two images line up when you set them side by side.

Four credits at 1K, eight at 2K and sixteen at 4K, identical on both engine tiers, so choosing the precise tier costs nothing extra. 1K generations are free to run; 2K and 4K output is part of Pro.