Enhanced Everything

This MCP Server turns the voidtools Everything search engine into a fast, AI-assisted development research platform. You can perform lightning-fast file searches, do intent-based code research, and drive spec-driven development with integrated tools and external assistants. It provides a practical bridge between quick local searches and in-depth project discovery across specifications, code, and documentation.

How to use

You connect an MCP client (such as Cursor, Kiro IDE, SpecStory, or Claude) to the Everything MCP Server and issue search requests through familiar commands. The server executes searches against Everything, formats results, and returns structured data that your client can display in its UI. Use it to quickly locate source files, documentation, specs, and related artifacts, then combine those results with AI-driven analysis and project context.

You can leverage built-in features like intent-based code research, cross-referencing between specifications and implementations, and project-wide knowledge extraction. The integration supports targeted searches, filtering by file types or dates, and sorting by name, size, date, or path. The system keeps detailed traces for debugging and auditing search activity.

How to install

Prerequisites you must have before starting: Windows OS, Python 3.8 or higher, a Python virtual environment, Everything installed from voidtools, Node.js 18 or higher, and es.exe accessible from the command line.

Step-by-step setup and run flow to get the MCP server up and running.

# Create and activate the Python virtual environment
python -m venv everything-mcp-env
everything-mcp-env\Scripts\activate

# Install Everything from voidtools.com (manual steps outside this snippet)
# Ensure the Everything service is running

# Ensure es.exe exists at the expected path
# Test Everything locally
"C:\\Program Files\\Everything\\es.exe" test

# Install MCP server dependencies and prepare for running the server
npm install

Additional notes

The MCP server is designed for local operation with standard inputs from MCP clients. The recommended workflow is to run the server in a Python-virtualized environment and start the Node-based MCP server process that exposes the search API to clients.

If you need to verify the Everything service, use a dedicated tool to check service health and perform trace logging for troubleshooting.

Troubleshooting

Common issues include missing es.exe, search results not returning, or permission problems. Ensure Everything is installed, the es.exe path is correct, and the Everything service is running with the necessary permissions.

Enable trace logging to diagnose search issues. The trace directory stores operation logs, development notes, and error traces.

Architecture overview

The server accepts requests from MCP clients, validates and processes them, executes searches via es.exe, and returns structured results. A trace system records operations for debugging and auditing.

Development and customization

You can build and run the server locally, modify the source, and test changes with the included scripts. The project includes a TypeScript source, built dist files, and a specification folder detailing enhancement plans.

Security and access

Limit access to the MCP server to trusted clients and ensure the Word/Windows permissions are set appropriately for es.exe and the Everything database. Use trace logs responsibly to avoid leaking sensitive data.

Examples

Use the Everything search to locate development files, spec documents, and API references, then hand off results to your AI assistants for context-aware analysis.

Available tools

everything_search

Basic file and folder search using Everything syntax to locate items quickly.

everything_search_advanced

Advanced search with extensive filtering, sorting, and regex support for precise results.

everything_check_service

Verify that the Everything service is running and accessible from the MCP server.

everything_search_docs

Specialized search for documentation and spec-related files optimized for SpecStory workflows.

everything_deep_research

Intent-based code research across public GitHub repositories to discover relevant implementations.

everything_github_search

Combined local and GitHub repository search to correlate code and documentation.

everything_tui_launch

Launch a terminal user interface for interactive project management and search workflows.