Concepts

Understand the fundamental concepts and architecture that power OpenClaw.

Core Concepts

• Agent - What is an OpenClaw agent
• Agent Loop - How the agent processes messages and executes actions
• Agent Workspace - The agent's file system and memory
• Architecture - System architecture and components
• Channel Routing - How messages are routed between channels
• Sessions - Managing conversation sessions and context
• Thinking & Reasoning - Agent decision-making processes
• Markdown Formatting - Cross-platform message formatting

Advanced Topics

• Subagents - Spawning specialized sub-agents for tasks
• Tool Use - How agents discover and use tools
• A2UI - Agent-to-UI communication protocol

Browse the topics above to deepen your understanding of how OpenClaw works.