Lesson 20: Agent Frameworks — OpenClaw, LangChain, and Others
The options for building an agent: from simplest to most complex. Which to choose based on your level.
Openclaw
Lesson 21: What is OpenClaw and Why We Chose It
We introduce OpenClaw, its philosophy, minimum requirements, and why it’s ideal for getting started.
Lesson 22: Installing OpenClaw Step by Step
Practical installation guide for Mac, Linux, and Windows (WSL). From zero to your first ‘hello’ with your agent.
Lesson 23: Configuring Your First Model and Provider
We connect OpenClaw to an LLM: choosing a model, getting a free API key, and configuring openclaw.json.
Lesson 24: Connecting OpenClaw to Telegram
Create a Telegram bot, connect it to OpenClaw, and talk to your agent from your phone.
Lesson 25: Giving Your Agent Personality — the System Prompt
Write an effective system prompt that defines how your agent behaves, speaks, and thinks.
Lesson 26: Essential Skills — Giving Your Agent Superpowers
The most useful skills to start with: web search, file reading, and real-time information sources.
Lesson 27: Automations — Letting the Agent Work on Its Own
Set up automatic tasks so your agent does things without you asking: briefings, alerts, monitoring.
Lesson 28: Multiple Agents — When One Isn't Enough
When it makes sense to have more than one agent, how to organise them, and how to combine expensive and cheap models intelligently.