Photo merger: two pictures into one

Turn on “Match brightness and color”. The tool measures each frame's mean brightness and saturation and pulls the others toward the first (which stays exactly as shot) or toward the middle of the set. The correction is capped at 0.8 of a stop: if the photos differ by more than that they really are different photos, and matching them all the way would be a lie. There is also a case where the tool deliberately does nothing — when the frames differ by a color cast rather than by light, say one under a lamp and one by a window, a single brightness cannot fix that and trying makes it worse, so the correction stands down on its own. The line under the switch says how many photos changed and how big the gap was.

Load exactly two photos and the panel offers it. The line is drawn on the join the layout engine computed, so it does not slide when you change the gap, the margin or the way sizes are matched. Labels sit on semi-transparent chips: plain white text on a bright photo is unreadable about half the time.

“Scale to match” resizes them to a common width or height keeping proportions; “Equal cells” gives an even grid, either cropping the overflow or fitting the whole frame; “Leave as they are” places the originals. Anything enlarged past its own resolution is flagged, and one button matches everything to the smaller photo instead.

Switch to “Layers”. The photo drags on the preview itself, the frame corners resize it and the grip above the frame rotates it; near an edge or the center it snaps and shows a guide. Opacity and 16 blend modes are in the panel on the right. This is how a watermark, a logo over a shot, and a double exposure are made.

Most likely “Styling” is set to a solid background. When any of the files you add carry transparency the tool switches the background to “Transparent” by itself on the first load — but if your previous merge used a white background, that preference was remembered from last time. Check the background switch and the format: the alpha channel survives a merge only in PNG and WebP, and choosing JPG raises a warning above the button that the transparent areas will be filled in.

When you capture a long page you grab the same rows twice, and every frame repeats the status bar and the bottom bar. The tool compares the frames, trims what repeats in all of them, then finds how far each pair overlaps and draws that part once. Every seam stays adjustable by hand.

Up to 40 at a time. Order them by dragging the thumbnails or with Alt and the arrow keys, and any photo can be rotated, flipped or removed — a removal comes back with “Undo”.

The first frame is used, and the tool says so. A still result cannot hold an animation, so nothing is lost silently.

Past a certain size the browser cannot hold the whole picture on one canvas, and an ordinary save would return an empty file. Instead the result is written row by row, and only PNG can be written that way — but it saves correctly at any length.