Browser extension for Chrome

The script tag works well for websites you can deploy to, but that's not always an option. Many of us also work on websites we don't control, like client sites, staging servers, store themes, or help centers.

A browser extension would let you open Pixy on any page with a keyboard shortcut after signing in once. You'd choose the project the page belongs to, and then edit it using the same tools and workflow you already know.

Nothing would be added to the website itself. The page visitors see would stay exactly the same, and strict content security policies wouldn't interfere because the extension runs in your browser.

Projects would work the same way they do today. You could connect a live site, staging server, and localhost to the same project, keeping all of your edits together regardless of where you made them. Your AI agent would receive the same changes it already does today.

Authorizing the extension creates a credential that belongs to your account rather than to a project or a website. Chrome stores it, and it never enters a page, so a site you edit cannot read it.

The script tag works differently. Its key sits in your site's source, where anyone reading the HTML can see which project the site belongs to. The extension puts nothing in the page at all. It identifies you, and you choose the project afterwards.

The extension remembers which project each website belongs to. It keeps one note per address, and Chrome syncs those notes to the other browsers you are signed into.

One project can hold your live domain, a docs subdomain, a staging server, and localhost. You bind each address once. If localhost points at a different site tomorrow, you bind it again and the rest of the project is unaffected.

The editor runs in the extension's own world. It sees the same page and the same DOM, but its JavaScript is separate from the site's.

That separation is what makes strict websites work. A content security policy controls what the page itself may load, and it does not reach the extension, so the editor opens on sites where a script tag would be blocked outright. It works in both directions: the website cannot see the editor or reach anything it holds.

⌘⇧E (Ctrl+Shift+E on Windows and Linux) is the only control the extension has. Press it on a page that already belongs to one of your projects and the editor opens.

Press it anywhere else and the popup asks which project the page belongs to, with a field to create one if the right project does not exist yet. That choice is the entire setup for a new website.

The editor ships inside the extension instead of being downloaded when you press the shortcut.

Chrome treats code fetched at runtime as a separate permission, and it puts that permission behind a toggle the person installing has to find on their own. Packaging the editor avoids the toggle. You install the extension, authorize it once, and press the shortcut.

The store listing is written. It says what the extension does in one sentence and gives a reason for every permission it asks for.

The build asks for three permissions and access to a single domain, which is Pixy's own. It does nothing on a website until you press the shortcut, and it reads nothing from the websites you never edit.

The build is in the Chrome Web Store's review queue.

Review times vary and the store gives no estimate, so there is no release date to publish yet. The extension itself is finished.

This is live now. The Pixy extension is on the Chrome Web Store: install it, authorize it once with your account, and press ⌘⇧E (Ctrl+Shift+E on Windows and Linux) on any website.

Chrome is the first one. Follow the browser extension entry for the others, which stays open while support for each browser is added.