n8n

n8n-MCP exposes n8n's documentation, properties, and operations to an AI assistant. This enables Claude and other models to understand and reason about 525+ n8n nodes, their schemas, and how to interact with them through available operations. The server runs in MCP stdio mode and streams responses back and forth with the host tool, providing access to node documentation, property schemas, and AI-capable node tooling. To use it, configure Claude Desktop (or your MCPlike client) with the provided mcpServers entry, then start the Docker container. You can opt to provide optional N8N API credentials to enable workflow management capabilities (create, update, execute workflows) in addition to the standard documentation tooling.

Prerequisites:

• Docker installed and running on your system (for the Docker-based option)
• Optional: Node.js if you prefer the local Node.js installation method

Option A – Docker (recommended for quickest setup):

1. Install Docker on your system (GitHub Desktop/desktop app, or package manager per your OS).
2. Run the MCP server (as shown in the mcp_config): docker run -i --rm -e MCP_MODE=stdio -e LOG_LEVEL=error -e DISABLE_CONSOLE_OUTPUT=true ghcr.io/czlonkowski/n8n-mcp:latest

Option B – Local installation (Node.js, for development or customization):

1. Prerequisites: Node.js >= 18 (check at https://nodejs.org/).
2. Clone the repository and install dependencies: git clone https://github.com/czlonkowski/n8n-mcp.git cd n8n-mcp npm install
3. Build and run locally: npm run build npm run rebuild npm start
4. Add a Claude Desktop config to point to the local MCP: { "mcpServers": { "n8n-mcp": { "command": "node", "args": ["/absolute/path/to/n8n-mcp/dist/mcp/index.js"], "env": { "MCP_MODE": "stdio", "LOG_LEVEL": "error", "DISABLE_CONSOLE_OUTPUT": "true" } } } }

Prerequisites recap:

• For Docker: Docker installed and running
• For Node: Node.js installed, cloned repo, dependencies installed

Note: Regardless of the method, you can optionally supply N8N_API_URL and N8N_API_KEY to enable workflow management features.

Tips and common considerations:

• The MCP runs in stdio mode; ensure your host tool communicates over stdio as configured.
• The Docker approach uses a minimal image with MCP runtime and a pre-built database; it contains no n8n dependencies to keep the image small.
• You can enable workflow management capabilities by providing N8N_API_URL and N8N_API_KEY; without credentials you’ll still have access to documentation and validation tools.
• Remember to set MCP_MODE=stdio and consider turning off console output for cleaner logs (DISABLE_CONSOLE_OUTPUT=true).
• If you use the Node.js local option, ensure you point the Claude config to the built dist/mcp/index.js path after building.
• Restart Claude Desktop after updating the MCP configuration so changes take effect.