Skip to content

Cookie settings

Optional analytics help us understand which pages and tools are useful. If you allow them, we use Google Analytics. Your files are never included, and every tool works the same if you decline. Cookie policy

Color Extractor from Image

Pull the colour palette out of a photo with the share each colour occupies, then copy the values as HEX, RGB or CSS custom properties.

Browser tool

Image

Extract the main colours of a picture. Analysis runs on your device.

About Color Extractor

Pulling a palette out of a photograph by eye tends to produce colours that are in the image but not representative of it. This counts the pixels, groups them, and reports the colours that actually dominate along with how much of the picture each one covers. Reading a palette this way is free and works online without a sign up: your photo is analysed inside the page, never uploaded, and the values you export carry no watermark.

How to use Color Extractor

  1. Open an imageJPG, PNG, WebP and GIF all work. The picture is read in your browser and is not uploaded.
  2. Read the palette and the sharesEach swatch comes with the percentage of the image it accounts for, which is what separates a background colour from an accent.
  3. Copy in the format you needHEX, RGB, or a block of CSS custom properties ready to paste into a stylesheet.

How the colours are grouped

A photograph of a "blue" sky contains thousands of distinct blues, so listing raw pixel values would return noise. Similar colours are clustered and each cluster is reported by a representative value, which is why the palette is short and stable rather than a list of near-identical swatches.

Transparency is excluded, not counted

Fully transparent pixels contribute nothing. Treating them as white — which some tools do implicitly — would report white as the dominant colour of a logo that has no white in it at all.

What it does not do

  • The palette describes the pixels, not the design intent. A photo shot in warm light returns warm colours even when the object photographed is neutral.
  • Very large images are sampled rather than measured pixel by pixel, so shares are close approximations and can move slightly between runs on the same picture.

Color Extractor: questions people ask

Why is the colour I can clearly see not in the palette?
Because it occupies very little of the image. A bright accent that catches your eye might be under one percent of the pixels, and the palette is ordered by area rather than by how much attention a colour attracts.
Can I use this to match a brand colour?
Only as a starting point. A colour sampled from a compressed photograph of a logo is not the brand value — JPEG compression alters colours near edges. Use the brand’s stated value where one exists.
Can I click on an image to pick one exact colour?
No, there is no eyedropper for single pixels here. The extractor groups all the pixels and lists the 2 to 16 main colours with the share of the image each covers, copyable as HEX, RGB or HSL. If you only need the one colour covering the most area, the Dominant Color Finder does that.

Questions about ToolsVerse itself — privacy, cost, what happens to your files — are answered in the frequently asked questions.

The tools people most often need alongside Color Extractor.

Can’t find the tool you need?

Tell us what you’re looking for and we’ll consider adding it.

Request a tool →