You have a photo and somewhere in it is the color — the cabinet paint, the sweater, the exact green of that café's wall. Getting from photo to usable code takes ten seconds; getting a code you can trust takes knowing three things about how photos lie.
Every real-world surface in a photo is at least three colors: the lit part, the shadowed part, and the highlight where light bounces straight into the lens. The paint's true color lives in the evenly-lit middle — never in the shine. Zoom mentally past the glossy spots and tap where the surface looks flattest and dullest. In the detector's photo mode the tap reads a small average rather than a single pixel, which shrugs off some noise, but it can't know you tapped a highlight.
JPEG compression stores color at half the resolution of brightness — the eye barely notices on photos, but it means the pixels along any sharp color edge are invented blends. Tap the middle of a colored area, at least a few pixels from any edge or text. Screenshots of interfaces are usually PNG and immune; WhatsApp-forwarded photos have been recompressed into soup, so get the original if the decision matters.
1 – For one precise color — a logo, a UI element, a product — use the detector: tap, read the name, copy the HEX.
2 – For "this photo's whole vibe," single taps mislead: sky that looks like one blue is forty blues. The palette extractor clusters every pixel and returns the six that dominate — the vibe, quantified.
3 – Either way, the codes are only the start: check text-worthiness with the contrast checker before the color ships anywhere near words.
The detector matches your pixel against 255 named colors in Lab space and tells you when the match is approximate — "closest name" with the exact HEX preserved. Any tool that always answers with a confident name is rounding your color to its nearest vocabulary word without telling you. The paint store doesn't care what the name is; they can work from the HEX. The name is for the conversation before that.