mine

A pi extension that adds a web_fetch tool for fetching web pages as clean markdown. It launches Chrome via Playwright on a virtual X display (Xvfb).

How it works

1. Opens the target URL in a headed Chrome instance
2. Waits for the page's load event (JS-heavy SPAs work fine)
3. Automatically dismisses cookie consent banners using I Still Don't Care About Cookies
4. Extracts the page content and converts it to clean markdown using Defuddle

Requirements

mine renders Chrome to a virtual X display so its window never appears on screen. Linux only. You need two things on your system:

• Google Chrome — Playwright launches the system Chrome (channel: "chrome"). Install from google.com/chrome or via your package manager.
• Xvfb — virtual framebuffer X server.

Distro	Xvfb install
Arch	sudo pacman -S xorg-server-xvfb
Debian/Ubuntu	sudo apt install xvfb
Fedora/RHEL	sudo dnf install xorg-x11-server-Xvfb

The extension manages the Xvfb process itself — start, display assignment, and shutdown are all automatic.

Install

pi install git:github.com/resolveworks/mine

For project-local install:

pi install git:github.com/resolveworks/mine -l

Usage

Once installed, the web_fetch tool is available to pi:

> Fetch the content of https://example.com

License

MIT