Skip to Content
About

About

NAICS MCP Server is an open-source project that brings AI-powered classification to the North American Industry Classification System (NAICS).

The Problem

Every business in America needs a NAICS code for government contracts, tax filings, and regulatory compliance. But the official lookup tools are stuck in 2003: keyword search against a static PDF, no semantic understanding, no context awareness.

The Solution

This project exposes NAICS classification through the Model Context Protocol (MCP) , enabling AI assistants to:

  • Semantic search across 2,125 industry codes using vector embeddings
  • Classify businesses from natural language descriptions
  • Navigate the hierarchy from 2-digit sectors to 6-digit national industries
  • Explore relationships between codes via pre-computed similarity graphs

Part of Cognition Kernel

This project is developed under Cognition Kernel , an independent research lab focused on AI safety, cognitive architectures, and practical AI tooling.

Other Cognition Kernel projects include:

  • Causal-Self - Python library implementing the Agentic Cognition Kernel framework
  • LazyDatabricks - Keyboard-first terminal UI for Databricks management

Author

Built by Fahad Baig .

License

MIT License. See the GitHub repository  for details.

NAICS Before the Internet