Course details
Openclaw
OpenClaw is an open-source gateway that connects chat apps such as Telegram and WhatsApp to an AI model you choose, on a machine you control. Learn how its Gateway, agents and memory fit together, where to run it safely, and how to install it, onboard it and keep it healthy.
Course curriculum
Meet OpenClaw3 lessons
OpenClaw is a self-hosted gateway between your chat apps and a model you choose. See how it differs from Claude Code, and sort real jobs by what starts them: a message, a clock or an event.
How OpenClaw Works5 lessons
One Gateway holds every channel and routes each message to an agent. The agent runs a loop of model calls and tool calls, guided by skills, while the context window and memory files decide what it can see and recall. Heartbeats and cron jobs let it act first.
Choosing Where It Runs4 lessons
Your own computer is quick for trying OpenClaw, a VPS gives uptime and reachability for a monthly fee, and dedicated hardware such as a Raspberry Pi or Mac Mini trades cloud bills for home maintenance. The security checklist applies to all three.
Installing and Onboarding4 lessons
Install OpenClaw with npm or in a Docker container, run the onboarding wizard to connect a model provider and a first channel, then set the Gateway's port, bind address and token and let a daemon keep it running.
Workspace Files and Health Checks4 lessons
SOUL.md and USER.md set who the agent is and who you are, MEMORY.md and the dated logs hold what it learns, and files like AGENTS.md, TOOLS.md, BOOT.md and HEARTBEAT.md each answer one question. doctor and security audit keep the whole setup working and safe.