Page Scanner

Save a browser tab into your vault as a vector PDF and a Markdown note. Scan a browser tab into the vault lists the tabs open in your browser, scans the one you pick, and puts two files in a folder of your vault: the PDF, with its text still selectable and searchable, and a note with the page's text as Markdown and the PDF embedded under its properties.

It works through Page Scanner, a Chrome and Edge extension that captures whole pages, and Page Scanner's command-line tool, which you install with npm. Desktop only.

What a scan leaves in the vault

In the folder set in the settings (Page Scanner by default), two files named after the page's title:

• Markdown - Wikipedia.pdf, the whole page as a vector PDF.

• Markdown - Wikipedia.md, the note:

  ---
  title: "Markdown - Wikipedia"
  source: "https://en.wikipedia.org/wiki/Markdown"
  captured: 2026-09-25T11:02:41.000Z
  ---
  
  ![[Page Scanner/Markdown - Wikipedia.pdf]]
  
  # Markdown
  
  Markdown is a lightweight markup language …
  

Characters a file name or a link cannot hold ([ ] | # ^ : / \) become dashes, and a name the vault already holds gets 2, 3 added.

Setting it up

1. Install Page Scanner 1.3.0 or newer from the Chrome Web Store or Edge Add-ons, in Chrome, Edge, Brave, Arc or Vivaldi.

2. Have Node.js 22 or newer on the computer, and install the command-line tool in a terminal:

   npm install -g @page-scanner/cli
   

   Obsidian cannot run it with its own runtime, so the plugin looks for it beside each Node it finds (Homebrew's, /usr/local/bin, Volta's, fnm's default, or your shell's) and runs it on the Node it was installed for. A Node can also be named in the plugin's settings.

3. In Obsidian, open Settings, Community plugins, Browse, search for Page Scanner, install it and enable it.

4. Open the plugin's settings. The steps at the top say what is done: press Install helper, then open the extension's settings (the gear in its popup), go to Local Agents and press Connect.

5. Run Scan a browser tab into the vault from the command palette, or press the ribbon's scan button, and pick the tab.

Settings

"As in Page Scanner's settings" leaves the choice to the extension's Capture settings. The extension editor's output (a crop, marks, a frame, a stamp, PDF/A) is not applied to a scan from Obsidian.

What it does outside the vault

• Files outside the vault. The plugin runs Page Scanner's command-line tool, the @page-scanner/cli package you install with npm, as a separate process with Node.js; to find it, it reads the global node_modules folder of each Node on the computer (and asks their npm where that is). The tool writes each scan into the system's temporary folder before the plugin moves it into the vault, and keeps its pairing and settings in ~/.page-scanner/. Install helper writes a small launcher there, and for each browser a file saying where the launcher is: in the browser's NativeMessagingHosts folder on macOS and Linux, and under a registry value in HKCU on Windows. That is how the extension and the tool find each other.
• Network use. None beyond this computer. The tool runs a small background process that listens on 127.0.0.1 only: port 45711 for the extension, and a second local port for the tool's own commands, both behind a token. The page is captured by your own browser, as it is already loaded there. Nothing is sent to Page Scanner or anyone else, and there is no telemetry.
• Nothing is downloaded or installed by the plugin. The command-line tool is yours to install and update with npm; the plugin only runs the copy it finds, and never fetches code.

License

Apache-2.0. The command-line tool it runs, @page-scanner/cli, is Apache-2.0 as well, by the same author.