Corrdex Reference

Corrdex is a local code intelligence engine built for structural code understanding, not just text search. It answers questions about architecture that grep and embeddings cannot — deterministically, from your actual source code.

This repository contains three user-facing surfaces.

Corrdex performs local static analysis and semantic classification across a project. It is not a text search tool or an embeddings layer — it uses explicit AST structure for every result.

Core capabilities

• Parse TypeScript, JavaScript, Python, SQL, and JSON
• Build a file dependency graph
• Classify files into architectural and semantic types
• Classify individual functions
• Detect workflow paths and runtime entrypoints
• Estimate change impact for files and functions
• Find hotspots, orchestrators, and architectural roles
• Run built-in rules plus custom policy rules
• Expose results through CLI and MCP

Typical outputs include

Corrdex is local-first. No remote inference, no embeddings for core understanding. Every result is derived from explicit structure in your source files.

High-level pipeline

What drives the analysis

• AST metadata extracted from source files
• Import resolution across the project
• Execution blocks and entry markers
• Call and member expressions
• File identities and naming heuristics
• Dependency categories
• Rule-based semantic inference

Current built-in parsing support covers the most common TypeScript and JavaScript ecosystems, plus Python, SQL, and JSON.

Install the published CLI/MCP package

This gives you both the corrdex and corrdex-mcp binaries globally.

npm install -g corrdex

Run from this repository

npm install
npm run build
node dist/cli/index.js --help

Build the npm package output

npm run build:npm-package

The publishable package lives under packages/npm.

From zero to a fully indexed, scanned project running MCP in five commands.

corrdex init .

corrdex index .

corrdex scan .

corrdex classify file src/orders/orderService.ts --explain

corrdex classify function src/orders/orderService.ts createOrder --explain

corrdex impact src/orders/orderService.ts

corrdex mcp .

corrdex-mcp .

After installing Corrdex, connect it to any coding agent that supports stdio MCP. Add a server entry in your agent's MCP config.

Standard (global npm binary)

{
  "mcpServers": {
    "corrdex": {
      "command": "corrdex-mcp",
      "args": ["C:\\path\\to\\project"]
    }
  }
}

If the agent cannot see global npm binaries (Windows)

{
  "mcpServers": {
    "corrdex": {
      "command": "C:\\Users\\HRD\\AppData\\Roaming\\npm\\corrdex-mcp.cmd",
      "args": ["C:\\path\\to\\project"]
    }
  }
}

Testing from repo before npm publish

{
  "mcpServers": {
    "corrdex": {
      "command": "node",
      "args": [
        "C:\\Users\\HRD\\code\\corrdex\\corrdex\\dist\\mcp\\server.js",
        "C:\\path\\to\\project"
      ]
    }
  }
}

The Corrdex CLI covers project initialization, indexing, scanning, cache management, file and function classification, semantic search, impact inspection, architecture summaries, and MCP server launching.

Main commands

Global flags

Hosted login and remote index

Use this when you want the local CLI to authenticate with hosted Corrdex and push one specific codebase.

corrdex login --path .

corrdex index . --remote

Corrdex exposes its local intelligence through MCP over stdio. The MCP server indexes the target project, hydrates an in-memory semantic registry, and serves deterministic tool outputs from the local project graph.

Startup behavior

• MCP connects immediately and does not wait for a full index before transport startup.
• Bootstraps from .corrdex-cache.json when available, then refreshes the full project index in the background.
• If first-time indexing is still running, call getIndexStatus.
• File-level tools can start from cache-backed state earlier than function-level tools.

Available MCP tools

Tool usage guidance

Prompts and resources

Recommended task memory shape

title: short task/topic name, preferably <area>: <outcome>
summary: 1-2 sentence outcome
details: compact sections like Decisions:, Constraints:, Verification:, Next steps:
tags: 2-5 stable keywords
files: only the touched or critical repo-relative paths

The VS Code extension runs a local language server, shows a Corrdex sidebar in the Explorer, provides a Corrdex AI view in the Secondary Sidebar, and refreshes on source and config saves.

Extension commands

Extension development

npm run build

The extension entrypoint is client/index.ts. The extension reads corrdex.config.json and corrdex.policies.json.

corrdex.config.json

Controls rule severities for local analysis.

{
  "rules": {
    "no-external-api-in-controller": "warning"
  }
}

Corrdex also accepts the legacy filename mergelens.config.json.

corrdex.policies.json

Adds compiled policy rules on top of built-in rules. Built-in rules are always loaded first — policy rules are appended after reading this file.

{
  "policies": []
}

Local cache

Corrdex stores semantic cache data in .corrdex-cache.json. The cache improves repeated indexing and scanning by reusing prior AST and classification work where possible.

corrdex cache status
corrdex cache clear

Optional remote upload

The local CLI works without any backend. If you want to push scan or index results to a backend, both commands support optional flags.

corrdex scan . --remote https://api.example.com --api-key key_xxx --project-id proj_123

Corrdex's local parsing, indexing, CLI, and MCP flow do not require remote infrastructure.

Corrdex produces structural semantic output at both file and function level. Results are deterministic — same source, same output.

Key directories

Important files