One Root, 63 Branches
The Claw ecosystem has grown far beyond its OpenClaw origin. From a single personal AI assistant, it has branched into 63 distinct projects spanning research automation, biomedicine, drug discovery, materials science, education, and multi-agent orchestration.
We curated every known project, classified its lineage, and built the interactive maps below. Click any project to open its GitHub repo.
OpenClaw Lineage (32 Projects)
Projects that are directly built on, inspired by, or deeply integrated with the OpenClaw codebase.
The deepest chain runs OpenClaw → NanoBot → PicoClaw → SciClaw — four generations of increasingly lightweight agents. Meanwhile, NanoClaw spawned a parallel chain: NanoClaw → MicroClaw → DrugClaw.
Highlights
- 6 direct descendants from OpenClaw core, written in TypeScript, Python, Rust, and Go
- NanoBot branch (5 projects) — the "lightweight AI assistant" lineage, from 35K-star NanoBot down to the 5-star SciClaw research agent
- NanoClaw branch (4 projects) — container-first security approach, with BioClaw adding bioinformatics on top
- 7 skill/domain packs — plug-in skill libraries that extend any OpenClaw-compatible agent
- 5 research systems — OpenClaw repurposed for academic workflows, from self-evolving colleagues to full research platforms
Independent Ecosystem (31 Projects)
These projects are not built on OpenClaw — they are standalone systems that share the "Claw" naming convention or operate in overlapping domains.
Highlights
- 3 alternative cores — FastClaw (Go), NullClaw (Zig), and ZeroClaw (Rust) offer OpenClaw-compatible alternatives
- 6 orchestration platforms — HiClaw (Alibaba), MetaClaw, OpenSpace, TinyAGI, Clawith, ArgusBot
- 18 research systems — the largest cluster, split across academic research, workspaces, and specialized tools
- AutoResearchClaw stands out with 7K+ stars and a 23-stage pipeline from idea to conference paper
Key Takeaways
- The NanoBot → PicoClaw → SciClaw chain shows how lightweight forks enable domain specialization at each generation.
- Research is the dominant use case — 23 of 63 projects target academic workflows.
- The independent ecosystem is nearly as large as the OpenClaw lineage (31 vs 32 projects), showing the "Claw" paradigm has become a category, not just a project.
- Cross-lineage orchestrators like HiClaw, MetaClaw, and OpenSpace treat all Claw variants as interchangeable backends — the ecosystem is converging on shared protocols.
Methodology
Each project was manually reviewed using its GitHub README, repo description, and visible code structure. Classification is based on:
- ancestor_l1: Root lineage (OpenClaw vs. Independent)
- ancestor_l2/l3: Category and subcategory
- relation_type: How the project relates to its ancestor
- confidence: High / Medium based on available evidence
The full dataset is available at claw-lineage-curated-sorted.csv.