Module 1 · Foundations
Run a useful agent with intentional model economics
Understand the system, install it, choose models, and keep recurring costs bounded.
You produce: A reproducible local agent and model-routing plan
Free curriculum · OpenClaw
Build an OpenClaw agent that runs 24/7.
A practical 16-lesson path from first run to a maintainable deployment.
16 lessons · 5 working outputs
Go from installation to an agent you can operate confidently.
Watch in order, apply one build step after every lesson, and leave with documented instructions, integrations, deployment, and recovery—not just a collection of videos.
Module 2 · Identity & Memory
Give the agent durable behavior and context
Connect a channel, define recurring behavior, and structure identity, memory, and workspace rules.
You produce: A channel-connected agent with reviewed instruction files
Module 3 · Tools & Integrations
Extend capability without losing control
Design narrow skills and connect external systems with explicit permissions and confirmation points.
You produce: One scoped capability and a least-privilege integration map
Module 4 · Deployment & Operations
Keep the agent running safely
Containerize, deploy, monitor, restart, update, and recover the agent predictably.
You produce: A repeatable deployment and recovery checklist
Module 5 · Advanced Workflows
Build a maintainable agent workspace
Use agentic coding, coordinate multiple agents, and organize the workspace for ongoing operation.
You produce: A verified workflow and complete workspace map
Get implementation feedback
Share redacted configs, architecture, and blockers with other builders in Kingstone AI Operators.
Join the free communityTurning this into an agency offer?
Diagnose the bottleneck between your technical build and your first AI-services client.
Find my starting point