Alqemist-labs

Alqemist-labs / agent_ferrum

Public

Browser automation library optimized for AI agents. Wraps Ferrum (Chrome headless via CDP) with compact snapshots: accessibility tree + markdown.

24
0
100% credibility
Found Feb 18, 2026 at 20 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Ruby
AI Summary

AgentFerrum is a Ruby library and command-line tool that automates headless web browsing for AI agents by providing compact, AI-friendly page snapshots with labeled interactive elements, stealth evasion, and download handling.

How It Works

1
📰 Discover the tool

You hear about a handy helper that lets smart assistants browse websites cleanly without all the clutter.

2
🚀 Launch your browser buddy

You start the simple command-line tool to open a private web browser ready for action.

3
🌐 Visit any website

You tell it to go to your favorite site, like a shopping page or news feed.

4
👁️ Get a clear page view

It shows you a neat summary of what's visible, with handy labels like @e1 on buttons and forms so your assistant understands easily.

5
🖱️ Click and interact

Using those labels, you click buttons, fill forms, or search just like a real person browsing.

✅ Mission accomplished

Your smart assistant finishes the job, like checking out items or gathering info, all smoothly and privately.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 20 to 24 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agent_ferrum?

AgentFerrum is a Ruby library for browser automation optimized for AI agents, wrapping Chrome headless via CDP with Ferrum. It generates compact page snapshots combining an accessibility tree of interactive elements (with clickable refs like @e1) and clean markdown of visible content, slashing token counts by 50-92% versus raw HTML. Developers get ref-based clicks, fills, and hovers without brittle selectors, plus smart waiting and download handling.

Why is it gaining traction?

It stands out in the github browser automation space by delivering AI-ready outputs that fit LLM context windows, unlike browser automation python or n8n tools dumping noisy DOM. The standalone CLI (agent_ferrum snapshot, click @e1) lets agents chain shell commands seamlessly, with stealth profiles evading bot detection. Benchmarks on GitHub repos and Wikipedia prove massive token savings for agent loops.

Who should use this?

Ruby devs building AI agents for web tasks like e-commerce scraping, form automation, or browser github copilot integrations. Ideal for agent_ferrum users chaining with browser automation mcp servers or n8n workflows, especially those frustrated by token limits in browser automation tools. Suits backend teams needing accessibility-focused snapshots without browser automation studio overhead.

Verdict

Try it for Ruby AI agent prototypes—excellent docs, CLI, and benchmarks make it production-ready despite 19 stars and 1.0% credibility score. Still early; watch for community growth before heavy reliance.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.