cartographer

<div align="center">

<br />

An agent-first code understanding system.

Build a persistent, evidence-grounded world-model of any codebase.<br /> Navigate it as a map. Inspect it down to source. Ask questions grounded in truth.

<br />

Getting Started · How It Works · Architecture · Contributing

<br /> <br />

</div>

---

You understand a codebase by building a mental model — boundaries, flows, state, invariants, failure modes. But that model lives only in your head. It's rebuilt from scratch every time. It can't be shared, inspected, or queried.

Cartographer externalizes that model.

It gives Claude Code a persistent memory of your codebase's structure and behavior — not as a static diagram, but as a living, evidence-grounded world-model that accumulates understanding over time.

You: "analyze this codebase"

Claude reads your code with its native tools (Read, Grep, Glob).
For each discovery, it writes a structured fact to the world-model:

  ┌──────────────────────────────────────────────────────┐
  │  entity:WorldModelStore                              │
  │  kind: entity                                        │
  │  confidence: proven                                  │
  │  evidence: src/store.ts:55-64                        │
  │  "Central state manager. In-memory Maps of           │
  │   WorldEntity and WorldRelationship, with debounced  │
  │   JSON persistence to .cartographer/model.json"      │
  └──────────────────────────────────────────────────────┘

Facts persist between conversations.
Open localhost:3847 to see the map.
Click any node to inspect source evidence.
Ask Claude questions — answers cite the stored model.

• Node.js 20+
• Claude Code CLI

In Claude Code, ru