A Chrome extension that copies any website — how it works, and why in-browser wins

A URL-based copier is a robot visiting a stranger's house. A browser extension is you, standing in the room, describing what you can see. That difference decides almost everything about fidelity.

What in-browser gets you that a URL cannot

• The page after JavaScript. Hero animations settled, lazy sections mounted, client-rendered routes populated. A fetch-and-parse tool gets the shell.
• Computed styles. Not the stylesheet as authored, but what the cascade resolved to for each node — including custom properties, container queries and media-query branches your viewport actually took.
• The right image. srcset and <picture> resolve per viewport and per DPR; the browser already picked, so the capture knows which file was used.
• Your own access. Pages behind your login, your paywall, your staging basic-auth, your geo — because you are already authenticated. Nothing leaves your machine to fetch it, and no credentials are handed to a third party.
• Canvas and WebGL. Painted pixels can be read out of the canvas; a server-side fetch has nothing to read.

The cost of that model is honest: one tab at a time, only pages you can open, and it runs on your machine rather than a server farm. For copying pages, that trade is overwhelmingly the right one.

What the extension does, step by step

1. Injects a capture pass into the active tab: walks the DOM, records computed styles per node, resolves pseudo-elements, reads canvas pixels, collects every asset URL the page referenced.
2. Fetches assets with the page's own credentials and referrer, paced per host so a site with strict rate limits (Wikimedia will 429 you happily) does not leave images behind.
3. Rebuilds the page as a self-contained document, then verifies itself — it re-renders its own output and compares box geometry and computed styles against the live page, and only ships the candidate that proves zero drift.
4. Exports standalone HTML, a ZIP, WordPress blocks, or an agent kit.

Step 3 is the unusual one and it is where the measured lead comes from: two candidate rebuilds are produced, one normalised and one keeping the original DOM verbatim, and the copy you get is whichever one re-renders identically.

Permissions, and why each is needed

Built on Manifest V3 with a service worker and a side panel:

• activeTab / scripting — to read the page you pressed capture on. Not "all sites, always".
• downloads — to hand you the exported file.
• sidePanel — the UI lives beside the page instead of in a popup that closes when you click away.
• Host access at capture time — to fetch the assets the page referenced, from the same origins the page itself used.

There is no analytics inside the extension, and captures are processed locally: the page content is not uploaded anywhere to be rebuilt. Licence checks and update checks talk to our own Worker and carry nothing about the page you captured.

Installing

The build is self-hosted rather than Web-Store-listed today, so installation is the standard unpacked route: download the ZIP from the download page, unzip it, open chrome://extensions, enable Developer mode, Load unpacked, pick the folder. Every release row on the download page carries its SHA-256 so you can verify what you unzipped. Updates are checked against the release API and announced in the panel.

FAQ

Is there a Chrome extension that can copy any website?

Yes — CopyAnySite is a Manifest V3 extension that captures any page you can open and exports it as standalone HTML, a ZIP, WordPress blocks or an agent kit. SingleFile is a good open-source alternative for single-file HTML only.

Does the extension work on pages behind a login?

Yes, because it captures the page already rendered in your session. Nothing is sent to a server to be fetched on your behalf, so no credentials are shared.

Does it work in Edge, Brave or other Chromium browsers?

It is built against Chromium's MV3 APIs and loads in Chromium-based browsers that support the side panel API. Firefox is not supported today.

Is the extension in the Chrome Web Store?

Not currently — it is distributed as a signed ZIP you load unpacked, with a published SHA-256 for each release.

Does the extension send my captures anywhere?

No. Capture and rebuild happen in your browser; only licence and update checks reach our server, and they contain no page data.

Keep reading

Clone any website — what a clone can and cannot include

Clone" promises more than any tool delivers. Here is the part of a website that genuinely clones, the part that never will, and how to tell which you need.

Copy any website's design — the layout, type and palette, not the content

How to extract a site's real design decisions — type scale, spacing rhythm, palette, grid, motion — and reuse them legitimately instead of copying the page itself.

Copy any website free — the options, and where each one breaks

Five genuinely free ways to copy any website, what each keeps, where each falls apart, and the measured fidelity of every one with the original server offline.

Copy any website's source code — and what "source code" really means

How to get the HTML, CSS and assets of any page you can open, why View Source usually shows you the wrong thing, and which parts of a site's code you can never have.