Color Picker from Image - Get the HEX Code of Any Pixel

Upload an image, point at any pixel, and read its exact HEX, RGB and HSL values. A pixel-grid magnifier shows what is under the crosshair, arrow keys nudge the sample one pixel at a time, and nothing leaves your browser.

What This Color Picker Does

This is an eyedropper, not a palette generator: you decide which pixel is sampled, and the tool reports exactly what is stored there. The part worth explaining is how a reading is taken, because the obvious approach is subtly wrong. Drawing the whole image onto one canvas and reading pixels back from it fails on large photos - browsers cap total canvas area at roughly 16.7 million pixels, about 4096 by 4096, and a 24-megapixel camera file exceeds that easily. Scaling the image down to fit would make every sample an average of several original pixels, so the one thing an eyedropper must get right would be the first thing lost. Instead, sampling redraws only a small window around the pointer, straight from the decoded file, at exactly 1 to 1. Source rectangle and destination rectangle are the same size, so the browser copies pixels rather than resampling them, and the values you get are the file's own bytes no matter how large the image is. The preview you see may be scaled to fit the canvas limit or your screen, but nothing is ever sampled from it. That same window feeds the magnifier, which paints each source pixel as a 13-pixel square with grid lines between them and outlines the centre cell in black and white so it stays visible over any color. You need it: a 4000-pixel-wide photo laid out in a 720-pixel column puts roughly 5.6 image pixels under every screen pixel. Arrow keys move the sample exactly one source pixel and Shift plus an arrow moves ten, which is the only reliable way to land on a one-pixel border or a thin antialiased edge. Each pick reports HEX, RGB and HSL from the same 8-bit triple, plus 8-digit HEXA and RGBA when the pixel is not fully opaque. The real limit to know is color management: canvas discards the embedded ICC profile in most browsers, so values match what the browser paints rather than what a color-managed editor shows.

Why Pick a Single Color from an Image

Most color questions are about one specific spot, not about an image as a whole. Someone hands you a screenshot of a signed-off design and you need the exact button fill to reproduce it in CSS. A client sends a JPG of their logo and you need the brand red before you can build anything around it. You are matching a chart series to the header color in a slide deck. You are checking whether two swatches that look identical in a photo really are the same value, or whether one is two steps darker. In all of those, an automatic palette is the wrong answer - it would return the six colors that cover the most area, and the color you actually need might be a 40-pixel accent that no clustering algorithm would ever surface. Being able to point at the pixel is the whole job. It is also the fastest route when the source file is gone. Design systems drift, source files get lost, and a flattened PNG export is often the only surviving record of what a color was. Reading it back takes a second and needs no design software installed. The same applies to work that never involved a designer: a photographer sampling a skin tone before grading, a developer copying a background color out of a bug report screenshot, a teacher pulling colors out of a painting for a worksheet. Because everything runs locally, unreleased mockups and internal screenshots never leave your machine.

Recover the exact fill of one button, badge or logo mark from a flattened screenshot when the source file is gone
Point at the pixel you care about instead of accepting whichever colors an automatic palette decides are dominant
Land on a one-pixel border or a thin antialiased edge using arrow-key nudging and a 13 by 13 pixel magnifier
Copy the same pixel as HEX, RGB or HSL, plus HEXA and RGBA when it carries transparency

How to Get a Color from an Image

Four steps, and no account or install. Files are read directly by the page - nothing is uploaded to a server.

1

Drop an image on the upload area, choose a file, or paste one straight from the clipboard. JPG, PNG, WebP, GIF and SVG all work.

2

Move the pointer over the image. The magnifier shows the pixel grid around the crosshair and the panel updates live. On a touch screen, press and drag - the magnifier sits above your finger so it is never hidden.

3

Fine-tune with the arrow keys or the on-screen arrows: one press moves the sample exactly one source pixel. Switch to actual size if the image is being scaled down to fit the column.

4

Click, tap or press Enter to keep the color, then copy its HEX, RGB or HSL value. Kept colors stay in the recent list so you can copy any of them again.

Frequently Asked Questions

Three things can pull the value apart. First, most browsers ignore an image's embedded ICC profile when it is drawn to a canvas, so a photo tagged Display P3 or Adobe RGB is read as if its numbers were sRGB - a color-managed editor like Photoshop applies the profile and reports something else. Second, your monitor renders sRGB through its own gamut and calibration, so the same value looks different on two screens. Third, a JPG's lossy compression genuinely changes pixel values near edges, so what looks like one flat area can hold several slightly different colors. The number here is the byte that is actually stored, which is what CSS and code will reproduce.
Canvas returns un-premultiplied color, meaning the RGB and the alpha are stored separately. A pixel authored as 50 percent white over nothing comes back as rgb(255, 255, 255) with alpha 128 - a value that never appears anywhere on screen. Whenever alpha is below 255 the tool shows the raw values, the 8-digit HEXA and RGBA forms, and additionally what that pixel resolves to over a white background, which is usually the number you actually wanted.
This page samples the pixel you point at. The palette page analyses the whole image, quantises it into clusters and returns the four to ten colors that dominate it, ranked by how much of the picture they cover. Use this one when you know which spot you care about - a logo mark, a button, one leaf. Use the palette page when you want a scheme derived from the image as a whole and do not have a specific pixel in mind.
No. The file is read by the page itself, decoded into a canvas in your tab, and sampled there. No copy is sent to a server, nothing is stored, and the recent-picks list lives only in the open tab and disappears when you close it. That matters for unreleased mockups, internal dashboards and anything else you would not paste into a random web form.
Press and drag rather than tapping. The magnifier is drawn above the contact point instead of under it, so you can see the pixel grid while your finger is still on the screen, and the sample follows the drag. For the last pixel or two, use the on-screen arrow pad: each press moves the sample exactly one source pixel, which no fingertip can do. Lift your finger or press the check button to keep the color.
Yes. Browsers limit total canvas area to roughly 16.7 million pixels - about 4096 by 4096 - so a 24 or 50 megapixel file cannot be displayed as one full-resolution canvas, and the preview is drawn smaller with a note saying so. Sampling does not go through that preview. Each reading redraws only a small window around the pointer, taken from the decoded file at 1 to 1, so the pixel you get is the pixel in the file. There is no need to crop the image first, and a single-pixel detail in a 50 megapixel photo reads exactly as authored rather than blurring into its neighbours.
Yes, with one caveat each. An animated GIF is drawn as its first frame only, so colors that appear later in the animation are not reachable - export the frame you need first. An SVG is rasterised before sampling, so you get the color of the rendered result including any opacity, filter or gradient applied to the shape, not the raw fill written in the markup. If the SVG declares no intrinsic size it is rasterised at 1024 pixels wide.
They all describe the same pixel, so it is about what you are pasting into. HEX is the compact form most design tools and stylesheets expect. RGB is easier to read when you need to nudge one channel or compare two colors numerically. HSL is the one to copy when you want a variation rather than the exact color: keep the hue and saturation, change the lightness, and you get a matching tint or shade. An 8-digit HEXA or an RGBA value appears as well whenever the pixel carries transparency.

その他の画像ツール

画像編集のニーズに応える他の強力なツールをご覧ください

画像圧縮

サイズ縮小

サイズ変更

寸法変更

画像切り抜き

領域を切り取る

形式変換

形式を変更

フォトエディター

写真を編集

画像回転

画像を回転

画像反転

水平/垂直反転

透かしを追加

透かし

枠線追加

画像枠線を追加

角丸

角丸を追加

フィルター

明るさ/コントラスト

顔をぼかす

顔を隠す

モザイク

領域をぼかす

背景削除

背景を削除

画質向上

AI拡大

テキスト追加

画像にテキストを追加

コラージュ

複数の画像を結合

HTMLから画像

Webを画像に

GIF からシーケンス

フレーム抽出

画像の余白を削除

空白の縁を削除

画像を一括リネーム

ファイル名を安全に一括変更

円形切り抜き

画像を円形に切り抜き

白黒

画像をグレースケールに変換

画像分割

画像をグリッドに分割

画像結合

画像を1つに結合

DPI変更

画像のDPIメタデータを設定

画像→Base64

画像をBase64にエンコード

画像をPDFに結合

複数の画像を1つのPDFにまとめる

HEIC→JPG

iPhoneのHEIC写真をJPGに変換

EXIF除去ツール

GPS位置情報を含む写真の隠しメタデータを表示・除去

ファビコン ジェネレーター

完全なアイコンセットとHTML

カラーパレット

画像から色を抽出

Base64 to Image

Decode Base64 back to an image

HEIC to PNG

Convert iPhone HEIC photos to lossless PNG

Text Behind Image

Put text behind the subject of a photo, entirely in your browser

Signature Maker

Draw a signature and download it as a transparent PNG

Signature Background Remover

Ink from paper, on transparency

Transparent Background Maker

Make a background transparent