How to think about connections — the difference between integration patterns (this section) and marketplace packages (§4 Plugins). Both matter; they answer different questions.
The integration patterns, not the marketplace packages. Channels bring messages in and out (Slack, WhatsApp, Discord, etc.). Tools let the agent do work (browser, filesystem, web search, code exec). Models are the brains (Claude, GPT, local Ollama). Memory is what survives a restart (built-in, Honcho, QMD). For specific plugins/packages reviewed individually, see §4 Plugins.
3.1 sourced 2026-05-07
3.2 sourced 2026-05-07
How messages flow in and out of OpenClaw — the 24+ messaging surfaces you can connect, what each one does well, and the security defaults that matter.
3.3 sourced 2026-05-07
What the agent can DO — built-in tools (read/exec/edit/write/browser/canvas/MCP) and how to think about tool policy.
3.4 sourced 2026-05-07
The brain. Which model providers OpenClaw supports, how to configure model refs, and how to think about model failover and cost.
3.5 sourced 2026-05-07
What the agent remembers. Three memory engines (built-in / Honcho / QMD), active vs long-term memory, and how to think about persistence.