Image Color Picker

Browsers may apply the image's color profile (sRGB conversion) on decode. For pixel-exact picks, save the source as plain sRGB and avoid color-managed formats.

Yes, but a fully transparent pixel picks as #000000, not as the colour you see behind it. The picker reads the raw canvas bitmap, where a pixel the image never painted is (0, 0, 0, 0), and the page behind the image is not part of that data. Even a PNG that stores a colour underneath its transparency loses it, because compositing discards RGB at zero alpha. Flatten the image onto the background you want before picking.

No. The picker uses Canvas getImageData locally. Cross-origin images loaded into the canvas would normally raise a security error, but uploads from your device are always same-origin.

It's a soft limit to keep the palette readable. Pick what you need, copy it, then clear and start again.

Because there is no averaging: the canvas is set to the image's full native size, your click position is scaled from the displayed size back to native pixels, and exactly one pixel is read. On a large photo shown small, one screen pixel covers several image pixels, so a one-screen-pixel move can land somewhere quite different. JPEG artefacts and camera noise make it worse. For a colour that represents an area, take several picks and average them, or blur the image first and pick from that.

The browser applies the image's embedded colour profile when it decodes, so a file tagged Display P3 or Adobe RGB is converted to sRGB before it ever reaches the canvas, while a design tool may show you the untagged original numbers. The picked value is what a browser will actually display, which is usually what you want for the web. If you need the file's own numbers, strip or convert the profile to sRGB before uploading.