llm-agents-radar

LLM Agents Radar is a self-updating tracker that discovers and indexes LLM agent frameworks appearing on GitHub. Rather than a hand-maintained awesome-list, it relies on a scheduled GitHub Action to keep its listing current. How it works: every 15 minutes a GitHub Action runs tracker.py, which fetches the latest state from the GitHub Search API, diffs it against the previous snapshot stored in data/items.json, rewrites the table between marker comments in the README, and commits a change summary when items are added or removed. The project states it uses no external services or paid APIs, only a public data source and a free GitHub Action. The README presents a table of roughly 50 tracked repositories with columns for name, star count, language, last update date and description. Entries span a wide range of agent-related projects, including terminal-based coding agents, multi-agent research terminals, memory providers, browser automation agents, deep-research tooling, agent governance and evaluation utilities, and various smaller experimental repos. The listing is regenerated on each cron tick, so the visible contents change over time. Contributions are described as usually unnecessary since the tracker maintains itself; issues are invited for data-source bugs or suggestions for new table columns. The project is released under the MIT license and links to a set of sibling trackers from the same maintainer covering adjacent topics such as MCP servers, Claude Skills, Cursor rules, RAG, LLM evaluation, vector databases and LLMOps.