Skip to content

Concepts Overview

Core building blocks that explain how Claude Code works and how to configure it.

Documents

Document Summary
Context window How Claude manages active information, compaction, and context size by model
Project What a Claude Code project is, memory hierarchy, team collaboration
MCP servers Connecting external tools and data sources via the Model Context Protocol
Tools Built-in tools, the agentic loop, permissions, and settings.json reference
Models Model aliases, switching, effort levels, extended context, prompt caching

How they relate

Model ──────────────────► sets reasoning depth + context window size
        │
        ▼
Context window ◄──────── Tools feed results back into it each step
        │
        ▼
Project (working dir) ◄── MCP servers extend what tools can reach
  • Model determines what Claude can reason about and how far back it can see.
  • Context window is the live workspace — tools write results into it, compaction trims it.
  • Project defines the directory scope and loads the memory layer on top.
  • MCP servers expand the tool surface beyond built-ins.
  • Tools are the actions Claude takes inside the agentic loop.

For persistent behavior configuration (memory, hooks, agents, skills) see Persistence.