Evidence before opinion.
Comparisons, incident reports, architecture, and benchmark methodology. Evergreen tutorials live in the handbook.
Best Ollama Models for OpenClaw: Tested for Tool Calling
A transparent OpenClaw + Ollama qualification guide: native API setup, minimum metadata, tool-call tests, scoring fields, and why we do not publish invented winners.
Read analysisOpenClaw vs Nanobot: Choose by Operating Model
A source-led comparison of OpenClaw and HKUDS nanobot across scope, language, channels, extensibility, operations, and the kind of owner each project fits.
Read analysisThe Moltbook Supabase Exposure: What Actually Failed
A source-led analysis of the 2026 Moltbook database exposure, why a public Supabase key was not the root cause by itself, and the controls agent platforms need.
Read analysisThe research shelf.
Model labBest Ollama Models for OpenClaw: Tested for Tool CallingA transparent OpenClaw + Ollama qualification guide: native API setup, minimum metadata, tool-call tests, scoring fields, and why we do not publish invented winners.Aug 9, 2026ArchitectureHow OpenClaw Works: Gateway, Agents, Tools & ChannelsA practical mental model of OpenClaw's Gateway, agent runtime, model providers, tools, Skills, plugins, channels, state, and diagnostic flow.Aug 9, 2026ComparisonOpenClaw vs Nanobot: Choose by Operating ModelA source-led comparison of OpenClaw and HKUDS nanobot across scope, language, channels, extensibility, operations, and the kind of owner each project fits.Aug 9, 2026Incident analysisThe Moltbook Supabase Exposure: What Actually FailedA source-led analysis of the 2026 Moltbook database exposure, why a public Supabase key was not the root cause by itself, and the controls agent platforms need.Aug 9, 2026