See how the tones of a photo are spread from shadows to highlights, per colour channel, with the numbers to match. Nothing is uploaded.
Counts every pixel in the photo and sorts it into 256 brightness bins for red, green, blue and luminance, then draws the result as a chart. The left edge is pure black, the right edge pure white. A tall pile on the left means a dark picture, a pile on the right a bright one, and a curve that stops short of either edge means the photo never reaches full black or white and could take more contrast. Below the chart you get the mean, median, spread and range of each channel, plus how many pixels are clipped at the ends.
A 4000 × 3000 photo of a beach at noon: the luminance mean comes out at 172, the median at 181, and 3.2% of pixels sit at 253 or above. That highlight clipping means the sky and sand have lost detail that no slider can bring back. The same scene at sunset reads mean 78, spread 41, no clipping: a dark image with room to lift. A screenshot of a white document shows a single spike near 250 and almost nothing else, which is what a histogram of mostly one tone looks like.
Use it to judge exposure before you edit: the chart shows whether a picture is dark, flat or blown out far more honestly than your screen does. Check it after a heavy edit too; gaps like a comb in the chart mean tones were stretched too far. It is not an editor. When the histogram tells you the photo is too dark, go to Change Brightness or Adjust Gamma; when it tells you the tones are bunched in the middle, Equalize Image or Change Contrast will spread them.
Channels switch the red, green, blue and luminance curves on and off; comparing the three colours shows a colour cast, since a neutral scene has them roughly aligned. Log scale compresses tall peaks so the small counts in the shadows and highlights stay visible. Filled or Lines is only a matter of taste. Show over the photo draws the chart onto the lower part of the picture, which is handy when you want to screenshot both together. The downloaded chart is a 1200 × 560 PNG.
What do mean and median tell me? The mean is the average brightness from 0 to 255; the median is the middle value. When the median is well below the mean, a few bright areas are pulling the average up and most of the photo is darker than it looks.
What counts as clipping? Pixels in the last three bins at either end (0 to 2 or 253 to 255). Above about 1% at one end, detail there is gone: pure black shadows or pure white skies.
Why does my histogram look like a comb? Gaps between thin bars appear when a low-quality or 8-bit image was stretched by a strong contrast or levels edit. The tones in between no longer exist in the file.
Is the photo uploaded to count the pixels? No. The counting runs in your browser with the canvas API. A 12-megapixel photo takes well under a second and never leaves your device.