16-Day Bootcamp
Go from zero to a fully autonomous AI agent running on your own server. Each lesson builds on the last — no code required.
What Is OpenClaw?
The complete beginner-to-advanced introduction. Understand what OpenClaw is, why it differs from every other AI tool, and what you'll build across this 16-day bootcamp.
Install & First Run
Complete installation guide for macOS, Linux, and Windows. Node.js, one-line install, API keys, first message, Telegram setup, and full troubleshooting.
Models & Ollama
Deep dive on every AI provider OpenClaw supports. Anthropic, OpenAI, Google, Chinese models, and full Ollama setup for running Llama 3 locally with zero API cost.
Cost Optimization
Two-tier model processing, heartbeat math, prompt bloat, spending limits, and prompt caching. The strategies that take a $70/month deployment to $17 with the same agent doing the same work.
Channels
Connect your agent to Telegram, WhatsApp, Discord, iMessage, Slack, and the built-in WebChat. One agent, every surface. Covers setup, security, formatting differences, and which channel to use when.
HEARTBEAT.md
Transform your agent from a chatbot into an autonomous system. HEARTBEAT.md defines what your agent checks on a schedule — morning briefings, task alerts, system monitoring — all delivered through your channels while you sleep.
SOUL.md
Your agent works, but it still sounds like generic AI. SOUL.md is where personality lives — tone, opinions, boundaries, style. Learn to write sharp behavioral directives, avoid corporate mush, and use the Molty prompt to craft a voice that's distinctly yours.
MEMORY.md
Deep dive into the three-file memory system: MEMORY.md for long-term facts, daily notes for running context, and the dreaming system that promotes what matters. Hybrid search, memory flush, token costs, and curation strategies.
AGENTS.md
Your agent has a voice and a memory — now it needs a rulebook. AGENTS.md defines operating rules, safety defaults, sub-agent delegation, tool policies, and the three security mechanisms that make autonomy safe.
Tools & Skills
Skills are modular packages that give your agent new capabilities. Learn the core built-in tools, install from ClawHub marketplace, build your first custom Skill, and understand the security risks of third-party code.
Integrations
Connect your agent to the real world. Messaging channels, webhooks, Google Workspace, web search backends, browser automation, and building custom integrations with the enterprise chain.
Docker
Move from bare metal to Docker. Isolation, reproducibility, sandbox modes, Docker Compose setup, security hardening, and production deployment patterns for running OpenClaw in containers.
VPS Deploy
Deploy OpenClaw to a VPS so it runs 24/7 without your laptop. Server selection, installation, daemon setup, remote access, security hardening, and monitoring for always-on agents.
Agentic Coding
OpenClaw isn't just a chatbot — it's a coding partner. The exec/read/write/edit toolkit, prompting strategies for code, build logs, planning workflows, and sub-agent patterns for development.
Multi-Agent
One gateway, many agents. Multi-agent architecture, routing precedence, channel bindings, agent-to-agent communication, per-agent security isolation, and running specialized agents for different roles.
Workspace Files
The finale. Every workspace file, every config key, every interaction — tied together in one definitive reference. Bootstrap order, session management, config reload, security audit, and the full architecture map.