Grayscale

Grayscale strips the colour from an image, leaving only the luminance of each pixel. The result keeps the structure and the tonal range of the original — highlights, midtones and shadows — while removing the information that told you which colour each part was.

Removing colour is often the fastest way to evaluate a composition. Without colour to draw the eye, the things that actually carry an image become obvious: whether the shapes read, whether the contrast is doing any work, whether the frame is balanced. Designers use it as a check before committing to a colour scheme for exactly that reason.

The conversion is a whole-image operation with no parameters, so it is applied uniformly rather than to a selected area. The output is written as a PNG, which keeps the conversion exact and preserves any transparency in the original. Because it is done on a canvas in your browser, the image is never transmitted anywhere.

Alongside the fixed scope, several controls you might look for are absent. There is no strength or mix setting, so a partial desaturation that keeps a trace of the original colour cannot be requested, and no channel is singled out — channel-mixing options of the kind a photographic filter set provides are not offered. The result cannot be tuned for contrast afterwards, and the page reports nothing about the conversion: no luminance histogram, no per-pixel values, and no before-and-after view to judge against. The output is raster only as well, so a vector source is flattened into pixels rather than having its colours rewritten, and very large inputs remain bounded by browser memory.