navegador 0.1.0

AST + knowledge graph context engine for AI coding agents

Navegador

AST + knowledge graph context engine for AI coding agents.

Navegador parses your codebase into a property graph and makes it queryable. AI coding agents can ask "what calls this function?", "what does this file depend on?", or "show me everything related to auth" — and get structured, precise answers instead of raw file dumps.

navegador — Spanish for navigator / sailor

---

Why

AI coding agents load context by reading raw files. They don't know what calls what, what depends on what, or which 5 functions out of 500 are actually relevant. Navegador builds a structured map — then exposes it via MCP so any agent can navigate your code with precision.

---

Quick start

pip install navegador

# Ingest your repo
navegador ingest ./myrepo

# Load context for a file
navegador context src/auth.py

# Search for a symbol
navegador search "get_user"

# Check graph stats
navegador stats

---

MCP integration

Add to your Claude / Cursor MCP config:

{
  "mcpServers": {
    "navegador": {
      "command": "navegador",
      "args": ["mcp", "--db", ".navegador/graph.db"]
    }
  }
}

Available MCP tools:

Tool	Description
ingest_repo	Parse and load a repo into the graph
load_file_context	All symbols in a file + their relationships
load_function_context	What a function calls and what calls it
load_class_context	Class methods, inheritance, subclasses
search_symbols	Fuzzy search for functions/classes by name
query_graph	Raw Cypher passthrough
graph_stats	Node and edge counts

---

Graph schema

Nodes: Repository · File · Module · Class · Function · Method · Variable · Import · Decorator

Edges: CONTAINS · DEFINES · IMPORTS · CALLS · INHERITS · REFERENCES · DEPENDS_ON

---

Storage

Navegador uses FalkorDB (property graph, Cypher queries).

Mode	Backend	When to use
Default	falkordblite (SQLite)	Local dev, zero infrastructure
Production	Redis + FalkorDB module	Shared / persistent deployments

from navegador.graph import GraphStore

# SQLite (default)
store = GraphStore.sqlite(".navegador/graph.db")

# Redis
store = GraphStore.redis("redis://localhost:6379")

---

Language support

Language	Status
Python	✅
TypeScript / JavaScript	✅
Go	✅
Rust	✅
Java	✅

---

License

MIT — CONFLICT LLC