Paste the address of a page and press Find images. The extractor reads the page the way a browser receives it and lists every image it refers to, each with a preview, its pixel size, its file type and the place on the page where it was found. Tick the ones you want, or none to take everything on screen, and download them as a ZIP. The Brand kit button collects just the logo, the icons and the social share images.
Where websites keep their images
Saving images by right-clicking only gets you what is drawn on screen, and many image downloaders only read the src of each <img> tag. A typical site spreads its images across far more places than that, and the extractor checks each of them:
- When an image comes in several sizes through
srcset, you get the largest one rather than the small copy a phone would load. - Lazy-loaded images often have a blank placeholder in
srcand the real address indata-src,data-lazy-srcor a similar attribute. The extractor reads those attributes and skips the placeholder. - Hero banners, patterns and some logos are CSS backgrounds set with
background-image. These are read from inline styles, from<style>blocks and from up to five of the page's stylesheets. On python.org, for example, 20 of the 29 images found came from its stylesheets. - Some icons never appear on the page at all: favicons, the Apple touch icon, the Safari pinned-tab icon, Windows tile images and the app icons listed in the site's web app manifest.
- The Open Graph and X (Twitter) card images are what shows up when the page is shared in a chat or a post.
- Logos and icons written straight into the HTML as
<svg>code have no file address at all. Each one is saved as its own .svg file. - The rest come from
<picture>sources, video poster frames, images embedded as data URIs, and the logo and images named in the page's structured data.
How the logo is found
The image marked Site logo is the strongest of these signals: the logo named in the page's Organization structured data, which is the one Google can show next to the site in search results; an image in the header that links to the home page; or an inline SVG in the header with "logo" in its class, label or title. Other images with "logo" in their file name, alt text or class are listed under Logos too. That is useful on pages with a row of client or partner logos, but it means not every logo in that group belongs to the site itself.
For the sharpest copy of a logo, take an SVG if there is one, because it scales to any size. Otherwise the largest Apple touch icon, usually 180 × 180 pixels, or a 512-pixel icon from the web app manifest are often bigger than the logo image in the header. The favicon.ico file can hold several sizes in one file, typically 16, 32 and 48 pixels.
What is in the ZIP
- Each image file is saved exactly as the website serves it, with no compression or conversion, so the quality is the original quality. Inline SVG code is saved as it appears in the page, with the SVG namespace added when it is missing so the file opens on its own.
- The file extension is set from the file's real contents. A file named
photo.pngthat is really a WebP image is saved asphoto.webp, so it opens in the right program. - Files are sorted into folders: logos, icons, social, images, backgrounds and svg.
- An
images.csvfile lists every image with its size, its alt text, where it was found and its original address. The same list can be downloaded on its own with the Image list (CSV) button.
The ZIP is put together in your browser. The site's server fetches each image from the website because browsers don't let one site read another site's files directly, then passes it straight to you without keeping a copy.
Checking your own site's images
Run the extractor on your own pages and it doubles as an image audit. The summary counts the <img> tags that have no alt attribute at all, which screen readers and Google Images depend on. An empty alt="" is not counted, because that is the correct way to mark a purely decorative image. Each card also shows the image's pixel size and where it came from, which makes it quick to check whether the page has a share image at all, to find photos uploaded far bigger than needed, and to spot old icons you forgot about. Large photos can be made smaller with the Image Compressor or the Image Resizer, a missing icon set can be made with the Favicon Generator, and the Open Graph Checker shows how the share image looks in a link preview.
Limits to know
- One page is scanned at a time. To collect images from several pages, scan each address.
- The extractor reads the HTML the server sends. Images that a script adds after the page loads, such as more results on an endless-scroll page, aren't in that HTML, so they are missed. The same goes for images inside iframes.
- Pages behind a login, and sites that block automated visitors, can't be read. You will see the error the site returned, such as 403.
- Only public addresses on the standard web ports can be opened. The first 3 MB of a page is read, with up to 500 images and 60 inline SVGs listed.
- Each image can be up to 20 MB, and one ZIP up to 400 MB. Download links stay valid for 3 hours; after that, scan the page again.
Using the images
Downloading an image doesn't give you the right to reuse it. Photos, illustrations and logos belong to whoever made them or to the company they represent. Publishing them on another site or in your own work usually needs the owner's permission or a licence that allows it.
