Navegador¶

The project knowledge graph for AI coding agents.

Navegador builds and maintains a queryable graph of your software project — combining static code analysis with human-curated business knowledge — so that AI coding agents always have precise, structured context instead of raw file dumps.

navegador — Spanish for navigator / sailor. It helps agents navigate your code.

Two layers, one graph¶

┌─────────────────────────────────────────────────────────────────┐
│  KNOWLEDGE LAYER                                                │
│  Concepts · Rules · Decisions · WikiPages · People · Domains   │
│                                                                 │
│         ↕  GOVERNS / IMPLEMENTS / DOCUMENTS / ANNOTATES        │
│                                                                 │
│  CODE LAYER                                                     │
│  Repository · File · Module · Class · Function · Method        │
│  Variable · Import · Decorator · (call graphs, hierarchies)    │
└─────────────────────────────────────────────────────────────────┘
              stored in FalkorDB  (SQLite local / Redis prod)

The code layer is populated automatically by ingesting source trees. Python and TypeScript are supported out of the box via tree-sitter. The knowledge layer is populated by manual curation (navegador add), GitHub wiki ingestion, and Planopticon output (meeting and video knowledge graphs).

Quick start¶

pip install navegador              # Python 3.12+ required
navegador ingest ./my-repo         # parse + index the codebase
navegador explain AuthService      # what is this thing?
navegador search "rate limit" --all  # search code + knowledge together

What goes in the graph¶

Layer	Node type	Populated by
Code	Repository, File, Module	`navegador ingest`
Code	Class, Function, Method	`navegador ingest` (tree-sitter AST)
Code	Decorator, Import, Variable	`navegador ingest`
Code	CALLS / INHERITS edges	`navegador ingest` (call graph analysis)
Knowledge	Concept, Domain	`navegador add concept` / `add domain`
Knowledge	Rule	`navegador add rule`
Knowledge	Decision	`navegador add decision`
Knowledge	Person	`navegador add person`
Knowledge	WikiPage	`navegador wiki ingest`
Knowledge	(any)	`navegador planopticon ingest`
Cross-layer	ANNOTATES, GOVERNS, IMPLEMENTS	`navegador annotate`

Agent integration¶

CLIMCPBootstrap

The simplest integration: call navegador explain or navegador context from any shell script or agent tool definition.

# get context for the file the agent just edited
navegador context src/auth/service.py --format json

# look up a function before editing it
navegador function validate_token --depth 2 --format json

# find everything annotated with a business concept
navegador concept PaymentProcessing --format json

Run navegador as a Model Context Protocol server. Configure it once in your agent settings and all navegador commands become callable tools with structured input/output.

{
  "mcpServers": {
    "navegador": {
      "command": "navegador",
      "args": ["mcp"]
    }
  }
}

See MCP Integration for the full tool list and per-agent config snippets.

One command to install navegador, ingest a repo, and wire the agent hook for your preferred AI coding assistant.

./bootstrap.sh --repo owner/repo --wiki --agent claude

Supports --agent claude, --agent gemini, and --agent openai. See Agent Hooks for what the hook does and how to configure it manually.

Navegador¶

Two layers, one graph¶

Quick start¶

What goes in the graph¶

Agent integration¶

License¶