# n8n > Build and validate n8n workflows with an agent that knows all 2,000+ n8n nodes — most-installed n8n MCP. [Canonical HTML page](https://top-mcps.com/mcp/n8n-mcp) · [server.json](https://top-mcps.com/mcp/n8n-mcp.json) · [methodology](https://top-mcps.com/about/methodology) ## Install ### Claude Desktop — `claude_desktop_config.json` Paste under mcpServers. Fully quit and reopen Claude after editing. ```json { "mcpServers": { "n8n-mcp": { "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### Claude Code — `CLI or .mcp.json` Run from your repo. Commit .mcp.json to share with your team. ```shell # export N8N_API_URL=https://n8n.example.com/api/v1 # export N8N_API_KEY=your-api-key claude mcp add n8n-mcp -- npx n8n-mcp ``` ### Cursor — `.cursor/mcp.json` Global path: ~/.cursor/mcp.json. Reload window after editing. ```json { "mcpServers": { "n8n-mcp": { "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### VS Code — `.vscode/mcp.json` VS Code uses the "servers" key (not "mcpServers"). ```jsonc { "servers": { "n8n-mcp": { "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### Windsurf — `~/.codeium/windsurf/mcp_config.json` Open via Cascade → hammer icon → Configure. ```json { "mcpServers": { "n8n-mcp": { "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### Cline — `cline_mcp_settings.json` Open via the Cline sidebar → MCP Servers → Edit. ```json { "mcpServers": { "n8n-mcp": { "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### Continue — `~/.continue/config.json` Continue uses modelContextProtocolServers with a transport block. ```json { "experimental": { "modelContextProtocolServers": [ { "transport": { "type": "stdio", "command": "npx", "args": [ "n8n-mcp" ], "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } ] } } ``` ### Codex CLI — `~/.codex/config.toml` Codex uses TOML. Each server is a [mcp_servers.] subtable. ```shell # ~/.codex/config.toml [mcp_servers.n8n-mcp] command = "npx" args = [ "n8n-mcp", ] env = { N8N_API_URL = "${N8N_API_URL}", N8N_API_KEY = "${N8N_API_KEY}" } ``` ### Zed — `~/.config/zed/settings.json` Zed calls them "context_servers". Settings live-reload on save. ```jsonc { "context_servers": { "n8n-mcp": { "command": { "path": "npx", "args": [ "n8n-mcp" ] }, "env": { "N8N_API_URL": "${N8N_API_URL}", "N8N_API_KEY": "${N8N_API_KEY}" } } } } ``` ### ChatGPT — `ChatGPT → Apps directory` n8n doesn't ship a hosted HTTPS endpoint today. ChatGPT supports remote MCP servers only — to use this server in ChatGPT you'll need to deploy it to a public HTTPS URL first (e.g. via Cloudflare Workers or Vercel) or wait for an official remote build. ```none ``` ## At a glance - **Maintainer:** czlonkowski (community) - **Transport:** stdio, Streamable HTTP - **Auth model:** None - **Required secrets:** N8N_API_URL, N8N_API_KEY - **Supported clients:** Claude, Cursor, VS Code, Windsurf, Any MCP-compatible client, n8n instances - **License:** MIT - **Language:** TypeScript - **Latest version:** latest - **Last verified:** 2026-07-16 - **Score:** 59/100 (rubric 2026-04 — see https://top-mcps.com/about/methodology) - **Source:** https://github.com/czlonkowski/n8n-mcp ## Quick answer **What it does.** Exposes n8n node documentation, property schemas, and workflow validation as MCP tools, plus optional workflow management (create, update, deploy) against a connected n8n instance. **Best for:** - Authoring n8n workflows with correct node config - Validating a workflow before import - Looking up node properties and operations - Deploying workflows via the n8n API - Working across 2,000+ core + community nodes **Not for:** - Hosted no-code automation (use Zapier/Make) - Simple trigger-only use - Teams not running n8n ## Description The most-installed n8n MCP (czlonkowski/n8n-mcp, MIT, 22k+ stars, ~107k weekly npm installs) gives an agent deep knowledge of n8n: documentation and properties for 2,000+ core and community nodes, workflow validation (including AI-agent nodes), and — with an API key — create/update/deploy against a live n8n instance. It is how you get an agent to author correct n8n workflows instead of guessing node schemas. ## Why it matters n8n has 2,000+ nodes; an LLM cannot hold their schemas in memory and hallucinates node config. Grounding the agent in the live node docs is the difference between a workflow that imports cleanly and one that errors on every node. ## Key features - Docs + schemas for 2,000+ n8n nodes - Workflow validation (incl. AI-agent nodes) - Optional workflow management via n8n API - Docs tools work with zero auth - npx + Docker, MIT ## FAQ ### Do the documentation tools need auth? No. Node documentation, property lookups, and workflow validation work with zero credentials — that is the default `npx n8n-mcp` mode. ### How do I manage workflows on my instance? Set `N8N_API_URL` and `N8N_API_KEY` to unlock create/update/deploy tools against a live n8n instance. Cloudflare Access is supported via `N8N_CF_CLIENT_ID`/`SECRET`. ### Is this the same as n8n's built-in MCP nodes? No — n8n also ships native MCP trigger/client nodes for using MCP inside workflows. This server is the external knowledge + build layer: it teaches the agent n8n's node catalog so it can author correct workflows. ## Changelog - **2026-07-16** — Refreshed install snippets and fact sheet; verified for 2026. - **2025-05-01** — Initial directory listing.