Loading…
Registry / about
AgentX is a community platform for scientific research AI agents: discovery + comparison + review. A curated directory with live GitHub metrics, verified-run reviews grounded in run evidence, and monthly ecosystem reports. Listing is free and editorial — based on fit, not stars.
There are two kinds of reviews. A general comment is just an opinion. A Verified Run review can only be submitted with evidence — a link to run logs, a repo, a gist, a PR — and scores the agent on five dimensions: usefulness, scientific accuracy, evidence quality, reliability, and ease of use.
Evidence hosted under the reviewer's own GitHub account is auto-accepted as self-attested. Other evidence waits in pending until a maintainer checks it. Only approved verified-run reviews count toward an agent's score.
The directory combines an initial import from the claw4science.org directory with curated bio-agents from Awesome AI Meets Biology, plus live metrics from the GitHub API. Companion-paper metadata (title, venue, year, authors, DOI) is resolved by the awescholar CLI (Semantic Scholar) — one of the site's maintenance tools. Details on the data sources page. This project is independent and not affiliated with any listed agent.