Quicky.Page

Browser extension

Publish exactly the part of a page you want. Select an area first, or send the full page, and you land in the editor with a private draft to review before making it live.

Availability

The extension is currently in development. Chrome Web Store and Firefox Add-ons listings are not yet live. Check back soon — or use the public API or MCP server to publish from a script, Claude Desktop, or Cursor in the meantime.

How it works

The extension only runs when you click. There is no background page scanner, no telemetry, no on-load injection.

  1. Optional: click Select area in the extension popup, then move your mouse over the page. The selected area stays highlighted with the rest of the page dimmed until you clear it or send it.
  2. Click the highlighted area, then click Send to Quicky.Page in the page overlay. Only the selected element is converted and sent. On chatgpt.com, chat.openai.com, or claude.aiit walks the message DOM and produces a clean chat transcript. Everywhere else it runs Mozilla's Readability — the same algorithm Firefox uses for Reader View — to grab the article's main content.
  3. It POSTs the result to /api/v1/publish as an unpublished page and opens the editor at quicky.page/?id=#edit= (the edit key lives in the URL fragment so it never reaches the server) in a new tab so you can review before publishing.

Selecting an area

The selector behaves like a lightweight inspector: hover to preview a content region, click to lock it, and use Clear from either the page overlay or the popup to go back to full-page publishing. The overlay is temporary UI injected into the current tab only after you click Select area.

Once an area is locked, the overlay shows its own Create new draft and Add to existing buttons. Choosing Add to existing opens a page picker on top of the current page, so selection stays first and the target page is chosen only when you decide to append.

The selected element is converted with the same limited HTML-to-markdown path used by the full-page extractor, so headings, paragraphs, lists, links, images, blockquotes, and basic emphasis survive. Scripts, forms, iframes, video, and other interactive chrome are ignored.

What permissions does it ask for?

  • activeTab + scripting — to extract content from the tab you click on. Only on click; no persistent access.
  • contextMenus — to add the right-click "Send to Quicky.Page" entry.
  • storage — local storage for the recent-publishes list inside the popup.
  • Host permission for https://quicky.page/*— to call the public API, check whether you're signed in, and show your saved pages in the popup.

That's it. No host permissions for chatgpt.com, claude.ai, or anywhere else — content extraction uses activeTab, granted at the moment you click.

History and accounts

When you're signed out, the popup keeps the same kind of local edit-link history as the website. After sending, it also writes that entry into the Quicky.Page editor tab so the website history menu and extension popup converge on the same pages for this browser.

You can also add selected content to an existing page. After selecting an area, choose Add to existingand pick a page from your local history or account list. Appending requires the page's edit key, so only pages you can edit are shown.

When you're signed in on quicky.page, the popup reads your account pages from /api/account/pages. New extension drafts are saved to your account when the browser session is available to the extension; if the browser blocks that session handoff, the extension falls back to local history and gives you first-party sign-in links.

What it does NOT do

  • Read pages you haven't clicked. activeTab means extraction is gated on a click, every time.
  • Send your data anywhere except quicky.page. The publish API call is the only request containing page content.
  • Track you. No analytics, no telemetry, no fingerprint collection.
  • Bring AI into the loop. The extension is a publishing primitive — it does not call /api/transform or any AI endpoint. If you want the content reshaped, do it inside the editor on the published page.

Building automation instead of using the extension? The same /api/v1/publish endpoint powers it. Want to publish from inside Claude or Cursor? See the MCP server.