A field guide to the AI web-agent zoo

What it does

This is an awesome-list that catalogs the rapidly expanding ecosystem of AI web agents. It groups projects into categories—autonomous browser agents, computer-use frameworks, automation dev tools, scrapers, and research benchmarks—so you can find the right project without drowning in GitHub search results. The list is maintained by the team behind Steel.dev, an open-source headless browser API built specifically for AI agents.

The interesting bit

The list doubles as competitive intelligence for its maintainers: Steel gets to map the territory while positioning its own browser API as infrastructure for the entire category. It is also unusually current, covering everything from OpenAI Operator and Anthropic Computer Use to experimental Chrome extensions and academic benchmarks.

Key highlights

• Covers the full stack: autonomous agents, computer-use frameworks, dev tools, scrapers, search tools, and research benchmarks.
• Juxtaposes heavyweight research prototypes (Google Project Mariner, Microsoft Magentic-One) with scrappy Chrome extensions (VimGPT, Nanobrowser).
• Maintained by Steel.dev, an open-source headless browser API project, suggesting the curation is dogfooded.
• Explicitly includes benchmarks and academic research alongside shipping products.
• The README itself admits when entries are thin—e.g., Comet by Perplexity has “not much more details out yet.”

Caveats

• The list is a marketing vehicle for Steel.dev; the “About Steel” banner dominates the top of the page.
• Some entries are placeholders rather than evaluations—descriptions occasionally amount to little more than a tagline and a link.
• The provided source is truncated, so coverage of later sections (scrapers, search tools, benchmarks) is incomplete.

Verdict

Worth bookmarking if you are building or evaluating browser agents and need a lay of the land. Skip it if you want deep technical comparisons or hands-on tutorials; this is a phone book, not a review.