What Color Is This? Color Name Identifier
Load a photo, click anywhere, and get the colour's name — with an honest measure of how confident that name is. Built for the moments when you need a second opinion on what you are looking at.
What this can and cannot knowRead before relying on an answer
Find out what colour this is
Load a photo and click anywhere in it, or type a colour value directly.
Click any point. A 7×7 pixel patch is averaged, so noise and compression artefacts do not decide the answer.
This colour is
coral
In the right family, but sitting between names — people would reasonably disagree.
It could also reasonably be called
Nearest CSS keyword: coral (ΔE 11.2). The headline answer is restricted to ordinary English, which is why the two can differ.
How this colour looks to other people
The name above does not change — the colour is what it is. This is what varies between viewers.
Typical
Protanopia
Deuteranopia
Tritanopia
Achromatopsia
Method
Why the Confidence Matters More Than the Name
Any tool can print a colour word. The useful part is knowing when not to trust it.
Most of this site is built for people making interfaces. This tool is built for the opposite situation: you cannot tell what colour something is, and you need to ask. That inverts what matters, because you cannot check the answer against what you see. A confidently wrong name is worse than no name.
So every answer carries the distance it was matched at. The colour you pick is converted to CIE L*a*b*, a space arranged so that equal numeric distances feel like roughly equal differences, and compared against a list of named colours. The nearest name is the answer; how near it landed is the confidence.
Where the names come from
They are the CSS named colours — the keyword list every browser implements, with fixed published values. That is a deliberately modest choice. There is no authority on what a colour “really is”; naming is cultural, continuous and varies by language. Using a fixed, checkable list means the answer is reproducible and you can look up exactly what was compared against.
The headline answer is restricted to ordinary English. The full list contains entries like papayawhip and gainsboro, which may sit numerically closer but are useless things to say out loud. The nearest full-list keyword is still reported underneath, so nothing is hidden.
Why a patch, not a pixel
Clicking samples a 7×7 patch and averages it. A single pixel is a bad witness: JPEG compression stores colour at reduced resolution, sensors add noise, and anti-aliasing blends edges. Averaging removes all three.
But averaging has its own failure, and it is a nastier one. Average across the boundary between a red shirt and a green wall and you get a confident brown that exists nowhere in the picture. So the spread within the patch is measured too, and when it is wide the tool says the spot is mixed instead of naming it.
| Label | Distance | What it means |
|---|---|---|
| Exact match | ΔE 0–2 | The colour is the named colour. No interpretation involved. |
| Close | ΔE 2–10 | Most people would use this name without hesitating. |
| Approximate | ΔE 10–25 | Right family, sitting between names. Two people would reasonably disagree — check the alternatives. |
| Rough guess | ΔE above 25 | Far from everything in the list. The alternatives are about as good as the headline answer. |
Limits
The Photo Is Not the Object
The single biggest source of wrong answers has nothing to do with the maths.
A camera does not record colour, it records light — and light is the object's colour multiplied by whatever is illuminating it. Under a warm bulb a white wall reflects orange light, and the file faithfully records orange. Your visual system corrects for this automatically and continuously, which is why the effect is invisible until a photo shows it to you.
Cameras attempt the same correction with automatic white balance, with mixed success. It can be fooled by a scene dominated by one colour, and phone cameras additionally apply saturation and contrast processing tuned to look pleasing rather than to be accurate. None of that is recoverable after the fact.
Practical consequences: photograph in daylight where you can, avoid coloured surfaces bouncing light onto your subject, keep out of shadow, and be sceptical of anything shot under a mix of daylight and indoor lighting. If a result matters, take two photos in different light and see whether the answer holds.
Screens add a second layer. If you are reading the swatch on an uncalibrated display with night mode active, what you see is shifted again — though the name and the numbers stay correct regardless, which is the point of showing hex and RGB alongside.
FAQ
Colour Identification Questions
What people ask about naming colours from a photo.
How accurate is this?
The maths is exact — the same hex always yields the same name. The uncertainty is in the photo, not the calculation. Under good daylight with a clean, flat subject, expect the name to be right. Under mixed indoor lighting, in shadow, or on a glossy surface, treat it as a hint. The confidence label reflects distance from the nearest name; it cannot tell you the photo was misleading, which is why the lighting advice matters more than the number.
Why does it sometimes give a name I would not use?
Because names are drawn from a fixed published list rather than invented, and colour naming genuinely varies between people and languages. Where a colour sits between two names the tool shows alternatives with their distances, which is a more honest answer than picking one and sounding certain. If the confidence reads approximate or rough, the alternatives are just as defensible.
Can I use it with my camera live?
Not yet — this version works with a photo you load or a colour value you type. Live camera capture is the natural next step and is planned. In the meantime, taking a photo first is actually more reliable: you can steady the shot, check the lighting, and re-sample the same image as many times as you like.
Is my photo uploaded anywhere?
No. The image is decoded into a canvas element inside your browser and the pixels are read back directly in JavaScript. It never touches a server, nothing is stored, and closing the tab discards it.
Why does clicking sometimes say the spot is mixed?
Because the 7×7 patch it sampled contained meaningfully different colours — usually an edge between two objects, a shadow boundary, or a patterned surface. Averaging those together produces a colour that is not actually present anywhere, so the tool declines to name it rather than inventing one. Click somewhere flatter.
What is the difference between this and the colour picker?
This answers “what colour is this?” from a photo. Our colour picker answers “how will this colour I have chosen be seen?” and suggests safer alternatives. One is for identifying, the other for designing. If you are comparing two colours rather than naming one, the pair checker is the right tool.
Does it work on a phone?
Yes. Choosing a photo opens your camera roll, and on most phones offers to take a picture directly. Tapping the image samples that point exactly as clicking does on desktop. Large photos are scaled down before sampling so the page stays responsive on mobile.