Installation

Get LORE MCP Server running on your machine in under a minute.

Quick Start: The fastest way to try LORE is with npx. No installation required.

Prerequisites

LORE requires Node.js 18 or later. We recommend using the latest LTS version for the best experience and compatibility with all analyzers. You can check your Node.js version by running node --version in your terminal.

While LORE is designed to work with any JavaScript or TypeScript project, it provides the richest analysis results for TypeScript codebases due to its type-aware analyzers. JavaScript projects still benefit from all structural analyzers including dependency graph, circular dependency detection, and entropy scoring.

Install via npm

For permanent installation, use npm to install LORE globally. This gives you access to the lore command from anywhere on your system, making it easy to run analysis on any project without navigating to a specific directory.

# Install globally
npm install -g lore-mcp

# Verify installation
lore --version

# Run analysis on your project
cd your-project
lore init
lore status

After installation, the lore CLI command becomes available globally. You can run it from any terminal session, on any project directory. The first time you run lore init on a project, it creates a .lore/ directory to store cached analysis results and configuration.

Try with npx (No Install)

If you want to try LORE without installing it permanently, use npx. This downloads and runs LORE on-the-fly without adding it to your global packages. It is the recommended approach for first-time users who want to evaluate LORE before committing to an installation.

# Initialize LORE on your project (no install needed)
npx lore-mcp init

# Run a quick status check
npx lore-mcp status

# Run the doctor for a health check
npx lore-mcp doctor

Tip: For MCP server usage with Claude Desktop or Cursor, you don't need to install globally. The MCP config uses npx -y lore-mcp which auto-installs on each session.

Install via Docker

For containerized environments or CI/CD pipelines, you can run LORE inside a Docker container. This ensures a consistent runtime environment and avoids any Node.js version conflicts with your host system. It is particularly useful for automated analysis in continuous integration workflows.

# Build the Docker image
FROM node:20-alpine
WORKDIR /app
RUN npm install -g lore-mcp
ENTRYPOINT ["lore"]

# Save as Dockerfile, then:
docker build -t lore-mcp .

# Run analysis on a mounted project
docker run -it --rm -v $(pwd):/project lore-mcp init /project
docker run -it --rm -v $(pwd):/project lore-mcp status /project

Verify Installation

After installing LORE, verify that everything is working correctly by running the version command. This confirms that the CLI is accessible and the package was installed properly. If you see the version number, you are ready to go.

$ lore --version
LORE v0.1.6 — Architectural Memory for AI Coding

$ lore status
  LORE v0.1.6 — Code Archaeology Report
  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

  Project:  my-project
  Files:    87
  Entropy:  0.28/1.0
  Score:    91/100

  All 13 analyzers passed ✓

Next Steps

Once LORE is installed and verified, you have several paths to explore depending on how you want to use the tool:

  • Explore Commands — Learn about init, status, doctor, diff, and more.
  • Set Up MCP — Connect LORE to Claude Desktop, Cursor, or Windsurf for AI-powered analysis.
  • See Examples — View real-world analysis output from popular frameworks like Express, NestJS, and Next.js.
  • Read FAQ — Get answers to common questions about LORE.