Free
5
pages per scan
- 1 website per account
- Up to 100 images optimized
- 200 MB for the results
- Files kept for 1 hour
- CSV and JSON maps, ZIP, agent prompts
No account needed.
Optimize every image on a website
Enter an address, a sitemap or a list of pages. Each page is analysed on its own, and you get optimized files in a ZIP with a map of what replaces what.
- Private by defaultDeleted within an hour
- Super FastProcess in seconds
- No Sign-upStart instantly
Every page, not just the home page
1. Each page is analysed on its own
From an address, a sitemap or your own list, the scan visits every page within your limit and records what that page loads: images, style sheets, scripts and fonts, and how the page uses them.
2. Each file is measured once
A logo used on 400 pages is downloaded once and listed with all 400. You see what is too heavy, too large for where it is shown, in an old format, or missing lazy loading, per page and for the whole site.
3. You get files and a map
Optimized images and minified style sheets at the same paths as the originals, a replacement map as CSV, Excel and JSON, a ZIP of everything, and ready instructions for Claude Code, Codex or any coding agent.
What we find is labelled as detected, what we predict as estimated, and what we produced as measured. Unused CSS and JavaScript are not measured, and scripts and fonts are analysed but never rewritten. Downloading the results changes nothing on your website until you put them in place.
Images are usually the heaviest part of a web page, and most sites still serve them as JPEG or PNG. This tool visits each page you point it at and records every image, style sheet, script and font the page loads. It converts the images to WebP or AVIF at a quality that keeps them looking the same. You get a table for every page and every file, with sizes before and after, so you can see exactly where the weight is.
It looks further than plain image tags. It reads srcset and picture sources, lazy-load attributes, CSS backgrounds, social preview images and icons. Start from an address and it follows the links of the same site, or give it a sitemap or your own list of pages. A logo used on every page is converted once and listed with the pages that use it. Vector images and anything that would not get smaller are left alone. A free scan covers up to 5 pages and 100 images, Pro up to 50 pages and 500 images, and a scan package a whole site of up to 50,000 pages.
The download is a ZIP with the optimized files at the same paths as on your site. It holds a replacement map as CSV and JSON, on Pro also as an Excel workbook, and ready instructions for AI coding assistants. Hand the map to a developer, or let an assistant apply it through our MCP server, which shows every change first and keeps a backup. Without an account the files are deleted after one hour. Pro keeps a scan for 7 days and a scan package for 30.
How many pages one scan covers
The limit counts for the whole scan, whichever way you start it: 50 pages on Pro means 50 pages in all, not 50 per sitemap. For whole websites there are scan packages: a number of pages to spend on any websites, in as many scans as you like.
Pro
50
pages per scan
- 5 websites per account
- Up to 500 images optimized
- 1 GB for the results
- Files kept for 7 days
- CSV, Excel and JSON maps, ZIP, agent prompts
Website 1K
1,000
pages in the package
$19, paid once
- Any website, not counted on your plan
- Up to 20,000 images optimized
- 5 GB for the results
- Files kept for 30 days
- CSV, Excel and JSON maps, ZIP, agent prompts
- Any websites, any number of scans, within 365 days
Not on sale yet. Tell us you are interested
Website 3K
3,000
pages in the package
$39, paid once
- Any website, not counted on your plan
- Up to 60,000 images optimized
- 10 GB for the results
- Files kept for 30 days
- CSV, Excel and JSON maps, ZIP, agent prompts
- Any websites, any number of scans, within 365 days
Not on sale yet. Tell us you are interested
Website 10K
10,000
pages in the package
$79, paid once
- Any website, not counted on your plan
- Up to 150,000 images optimized
- 20 GB for the results
- Files kept for 30 days
- CSV, Excel and JSON maps, ZIP, agent prompts
- Any websites, any number of scans, within 365 days
Not on sale yet. Tell us you are interested
Website 50K
50,000
pages in the package
$199, paid once
- Any website, not counted on your plan
- Up to 300,000 images optimized
- 40 GB for the results
- Files kept for 30 days
- CSV, Excel and JSON maps, ZIP, agent prompts
- Any websites, any number of scans, within 365 days
Not on sale yet. Tell us you are interested
A package is a number of pages. Every page a scan analyses takes one from it, on any website and in as many scans as you like, until the pages are used or 365 days have passed since buying. Several scans can draw on one package at the same time. Pages that could not be loaded or were left out do not count, and a page is counted once however often its scan is retried; optimizing a scan's files again and downloading them again cost nothing more. Scanning a website again counts its pages again. Generating replacement images with AI is not part of any plan yet.
Putting the converted images to work
The ZIP is half the job; the other half is making your pages point at the new files. Upload the WebP files to the server first, beside the originals, so nothing breaks while you work. Then go through the replacement map. Each row gives the address a page uses today and the file that should take its place, and in most projects that is a search and replace in the templates or the content database. Leave the originals where they are until every reference has been changed and checked.
Afterwards, load the pages in a private window and run the scan again. Images that have been switched are now reported as already small enough, and what remains is your to-do list.
What a scan cannot see
The tool reads your pages as a visitor without an account would receive them. Anything behind a login, a shopping cart or a form is out of reach. It reads the HTML and the styles written in it, and does not run the page in a browser, so pictures that a script inserts after loading can be missed. It also obeys your robots.txt file: pages you have closed to crawlers are not visited.
The scan is also read-only. It never writes to your site and needs no password, plug-in or file access, which means it is safe to try on a live shop in the middle of the day.
Frequently asked questions
Does this change my website?
No. It only reads the page, the way a visitor's browser does, and gives you new files. Nothing on your site changes until you upload the WebP files and update the references yourself.
Which images does it find?
Images in img tags, srcset and picture sources, common lazy-load attributes, CSS backgrounds in the page and in its style sheets, social preview images, icons and video posters. Images that a script inserts after the page has loaded are not found.
Why were some images skipped?
SVG files are vectors and already as small as they should be. Other images are skipped when the WebP version would not be smaller than the original, when a file is over 20 MB, or when it could not be downloaded.
How many pages does it scan?
From an address it follows links to other pages on the same site. From a sitemap or a list it scans the pages named there. A free scan covers up to 5 pages and 100 images, Pro up to 50 pages and 500 images, and scan packages up to 50,000 pages. Only public pages work, because we fetch them from our servers without cookies, and we respect robots.txt.
How much faster will my site be?
The table shows the bytes saved, which is the part we can measure. The effect on loading time depends on your hosting and your visitors' connections, so test a page before and after with a speed tool.
Do I have to replace every image at once?
No. Start with the largest files in the table. On most sites a handful of images accounts for most of the weight.