@1337xcode/cortex

Code intelligence for AI agents. Indexes your repo into a call graph, exposes it over MCP. Agents query the graph instead of reading files.

Install

npx @1337xcode/cortex@latest install

Downloads the binary for your platform, verifies the SHA256 checksum, removes any stale cortex binaries from other locations, and drops it in your PATH. Works on Linux, macOS (Intel + Apple Silicon), and Windows. Zero-touch: one command, no manual cleanup needed.

What it does

Your AI agent asks "what calls processOrder?" and gets a 200-token graph result instead of burning 20,000 tokens reading files. That's the pitch.

Cortex parses your code with tree-sitter (29 languages), extracts functions/classes/imports/call edges, stores everything in SQLite. Agents connect via MCP and query the graph.

Quick start

cortex index     # parse repo, build the graph
cortex install   # detect your IDE, write MCP config
cortex serve     # start the MCP server

Numbers

• 29 languages via tree-sitter
• 32 MCP tools (or 5 in smart mode via the ask meta-tool)
• 127 files indexed in 535ms, incremental re-index in 13ms
• 10-100x token reduction on structural questions

What else

• Taint flow analysis, OWASP Top 10 detection, SBOM generation
• Cross-session memory that marks observations stale when code changes
• Multi-repo federation for querying across repositories
• 3D interactive graph visualization (cortex viz --export graph.html)
• CI quality gates with configurable thresholds
• Git hotspot analysis, coverage gap detection
• Works offline. No cloud, no API keys, no Docker.

MCP config

{
  "mcpServers": {
    "cortex": {
      "command": "npx",
      "args": ["@1337xcode/cortex", "serve"]
    }
  }
}

Security

Single binary. No network access. No telemetry. No env var harvesting. SHA256 verified on install. Full source at github.com/1337Xcode/cortex.

Docs

1337xcode.github.io/cortex

License

MIT