Mapping the LLM agent gold rush

What it does This repository is a curated awesome-list that catalogs open-source projects building autonomous and multi-agent systems on top of large language models. It sorts entries into taxonomic buckets—single-agent task solvers, multi-agent orchestrators, society simulations, frameworks, benchmarks, and platform APIs—acting as a directory for a crowded and fast-moving corner of open-source software.

The interesting bit The maintainer groups projects by architectural ambition rather than dumping them alphabetically, ranging from simple browser automators like Agent-E to multi-agent runtimes like Octopal and even agent-society simulators. The list also reserves space for less-hyped categories like benchmarking tools and academic paper lists, which most “awesome” lists skip.

Key highlights

• Covers the full stack: end-user applications, underlying frameworks, evaluation benchmarks, and platform APIs
• Includes niche experiments like Anima-i (studying memory inheritance across agent generations) and InkOS (a 10-agent novel-writing pipeline)
• Tracks both heavyweight frameworks (AutoGPT, OpenDevin) and lightweight utilities (OpenPaw, KodeAgent)
• Maintains sections for surveys and paper-list repos, linking implementations back to research
• Uses GitHub star badges inline, so popularity is visible at a glance

Caveats

• Descriptions are largely the projects’ own README taglines, offering little independent analysis or architectural comparison
• No explicit inclusion criteria or quality bar is stated
• The list is sprawling, and finding the right tool for your stack still requires clicking through to each repository

Verdict Worth bookmarking if you are surveying the LLM agent landscape or choosing a framework, but skip it if you are looking for a single tool to install—this is a map, not a destination.