How to Use This Tool
Type a hex value or a colour name. The nearest names are ranked by perceived difference, with the straight RGB answer alongside for comparison.
Nearest needs a definition
The obvious approach treats a colour as a point in RGB space and takes the straight-line distance. It is fast and it does not match perception, because equal steps in R, G and B are not equally visible — the eye is far more sensitive to changes in green than in blue.
Measuring in CIE Lab instead gives a space designed so that equal distances look about equally different. The resulting figure is ΔE, and the conventional reading is:
- Under 1 — not perceptible.
- 1 to 2.3 — perceptible only in a direct side-by-side comparison. The 2.3 threshold is the classic "just noticeable difference".
- 2 to 10 — noticeable.
- Above 50 — different colours by any reasonable account.
The two methods often pick different names, and the Lab answer is the one that looks right when you see the swatches together. Where they disagree, this page says so.
Where the values come from
Rather than shipping a table of 148 hex values, every name here is resolved by the browser's own CSS parser and read back as a normalised value. That makes the table authoritative by construction, and a name the browser does not recognise is simply dropped.
It is a small thing with a real benefit: a hand-typed colour table can contain one wrong digit that nobody notices for years, because the wrong colour still looks like a colour.
The list is a museum
The 148 names accumulated rather than being designed, and it shows:
darkgrayis lighter thangray— 169 against 128. The X11 and HTML colour lists disagreed when they were merged, and the inconsistency is permanent because pages depend on the current values.- Both spellings work for grey, in every compound:
darkgrey,slategreyand the rest are all valid. rebeccapurplewas added in 2014 in memory of the daughter of a web standards contributor. It is#663399.- Several names are near-duplicates —
aquaandcyanare the same colour, as arefuchsiaandmagenta.
When to use a name
Names are convenient in a prototype, a diagram, or a quick test, and they are the wrong tool in a
design system. There are only 148 of them, they cluster oddly, and they carry no relationship to each
other — darkblue is not blue darkened by any consistent amount.
A design system wants a scale defined deliberately, most easily in a perceptual space such as OKLCH where "the same colour, two steps lighter" is a calculation rather than a search through a century-old list.
