Log in

Launch Your OpenClaw from Scratch

Learn how to plan, launch, and refine your OpenClaw agent.

  • Beginner
  • 11 lessons
  • 2 hr 23 min
  • Taught by Michael Shimeles
Watch with Pro

What you'll learn

  • Understand Agent Basics Explain how LLMs use tokens, embeddings, context, and prediction so you can design clearer prompts and workflows.
  • Prompt with Precision Write concise instructions that define the task, constraints, process, and definition of done for reliable agent behavior.
  • Set Up OpenClaw Deploy OpenClaw, connect model access and communication, and configure the core files that shape identity, tools, memory, and routines.
  • Connect Practical Tools Connect services such as Notion, Trello, Google Calendar, search, and email, then validate each tool with real actions.
  • Automate Workflows Build skills, subagents, and scheduled automations that preserve context and handle repeatable work without blocking your main agent.
  • Refine and Scale Improve OpenClaw over time with recursive skill refinement, reverse prompting, memory, and lean context management.

Lessons 11 lessons · 2 hr 23 min

Every lesson comes with Pro, $29 a month.

  1. Lesson 1: Introduction to LLMs, Agents, and Context 2 min
  2. Lesson 2: How LLMs Work: Tokens, Embeddings, and Prediction 13 min
  3. Lesson 3: Why Prompts Matter and What Reasoning Models Do 10 min
  4. Lesson 4: Agents, Tools, and the Action Loop 15 min
  5. Lesson 5: Understanding Context Windows and Bloat 15 min
  6. Lesson 6: Skills and Progressive Disclosure 15 min
  7. Lesson 7: OpenClaw Setup and Core Configuration 20 min
  8. Lesson 8: Connecting Tools with Composio, Notion, and Trello 20 min
  9. Lesson 9: Subagents, Cron Jobs, and Workflow Automation 15 min
  10. Lesson 10: OpenAI Codex Setup and Subscription Integration 10 min
  11. Lesson 11: Recursive Skill Building, Reverse Prompting, and Memory 8 min

Your teacher

Michael Shimeles

DevX Engineer @ Convex

Michael Shimeles is a Toronto-based full-stack engineer and AI builder. He has experience building real automation systems and agent-oriented tooling. Michael focuses on mental models, practical prompt strategy, and agent orchestration so builders can move beyond experimentation to systems that work reliably.

You’ll learn from his direct experience integrating agents like OpenClaw into workflows and architecting recursive improvement structures.

Part of AI automation and agents This recording is one step on the automation path. See the rest in order.

Related recordings

  1. AI Automations (New & Improved)

    Build reliable AI automations with n8n, Claude, and structured outputs to classify data, orchestrate multi-step workflows, debug integrations, and ship practical systems that save time.

    Nate Grahek Intermediate · 20 lessons · 3 hr 42 min

  2. The Human-First Agentic Content Workflow Workshop

    Build human-in-the-loop n8n workflows that turn existing team knowledge into polished content assets while preserving brand judgment, approvals, and scalable publishing paths.

    Nate Grahek Beginner · 7 lessons · 1 hr 2 min