Concept recap
You have built a working community agent. Until now, each page showed you one step at a time. This lesson steps back and shows you the whole picture.
Nothing new gets installed here, and you don't have to change anything. Some pages ask you to open a file and look inside, so keep your terminal and Telegram nearby.
The topics
Each topic has its own page:
| Topic | What you learn | Why |
|---|---|---|
| What we did | Every lesson in one page, and what each one added to Louis. | So the steps become one story. |
| How it fits together | What a model, a harness, a tool call and data are, what NanoClaw does as the orchestrator, and how they worked together every time you talked to Louis. | These four ideas explain every AI agent, not only this one. |
| Inside NanoClaw | Where your personality change and your scheduled message were saved, and how to find them yourself. | Nothing about Louis is magic. It is all files you can read. |
| Our modified NanoClaw | That the NanoClaw in this course is our customised copy, what we changed, and how Claude changed it using skills. | A standard NanoClaw behaves differently, and you can customise yours the same way. |
| Where to go next | How to keep an agent running for real, and what to learn after this course. | Your laptop is a good place to learn, and a poor place to run a community's agent. |
Next
Start with What we did.