Overview

Endor integrates with AI tools using the Model Context Protocol (MCP). This enables AI agents to run isolated services and recreate development environments instantly.

Endor provides a foundation layer for AI tools, providing isolated services that run in seconds.

Quick Setup

After installing Endor CLI, run it in MCP mode:

endor mcp

Endor will start listening using the STDIO transport.

You can pass the --allow-net option to enable public network access in the running services. By default, services are fully isolated and they cannot connect to public networks. Read more on Networking.

endor mcp --allow-net

AI Tool Configuration

Choose your AI tool for detailed setup instructions:

Claude Code

Configure Endor with Claude Code CLI for terminal-based AI development.

Setup Guide →

Goose CLI

Connect Endor with Goose CLI for local AI agent development.

Setup Guide →

Cursor

Integrate Endor with Cursor IDE’s Composer Agent for seamless development.

Setup Guide →

VSCode

Enable Endor in Visual Studio Code with built-in MCP management.

Setup Guide →

Windsurf

Set up Endor with Windsurf IDE for team-based AI development.

Setup Guide →

What AI Agents Can Do

Once configured, your AI agents can:

• Start services instantly: MariaDB, PostgreSQL, Redis, Valkey, and more
• Create isolated environments: Each service runs in its own sandbox
• Run commands: run any Linux command in isolated environments
• Test integrations: Spin up services for testing without local installation

AI agents automatically identify when to use Endor based on your development needs, making database and service management seamless.