How the parts interact in one turn

Walk through a single user request to see the parts cooperate:

1. The request arrives. The surrounding system authenticates the user and authorizes what they can touch.
2. The orchestrator assembles context: it loads relevant memory (history, available artifacts), builds the system prompt, and gathers the tools available on this surface.
3. The orchestrator calls the model with that context.
4. The model streams reasoning and text, and may request tools. The orchestrator runs each tool (these may touch the surrounding system: storage, external APIs), and feeds results back into the loop.
5. The model produces a final answer. The orchestrator streams it to the user and writes the outcome into memory.

Every chapter in this folder is, in effect, a deep look at one of these steps.