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*

[](https://github.com/ConflictHQ/navegador/actions/workflows/ci.yml) [](https://pypi.org/project/navegador/) [](LICENSE) [](https://navegador.dev)

---

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.

---

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

---

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_