Workflow Automation★★★★★5/5
n8n-MCP
n8n-MCP gives AI assistants comprehensive knowledge of the entire n8n node ecosystem — 537 core nodes and 547 community nodes, with 99% property coverage and 2,709 workflow templates.
Instead of the AI guessing at node configurations (and getting them wrong), it can look up exact property names, required fields, and real-world examples from 2,646 pre-extracted configurations. It even validates AI Agent node setups, checking memory connections and tool wiring.
Available as npx, Docker, Railway deployment, or a hosted service with a free tier (100 calls/day). With 13.6k stars, it's the go-to tool for anyone building n8n workflows with AI assistance.
Pros
- + 99% node property coverage across 1,084 nodes
- + 2,646 real-world configuration examples from popular templates
- + AI workflow validation for Agent nodes, memory, and tools
- + Multiple deployment options: npx, Docker, Railway, hosted
- + 2,883 tests — well-tested codebase
Cons
- - Large dataset (~280MB Docker image)
- - Free hosted tier limited to 100 calls/day
- - Full workflow management requires N8N_API_KEY
How We Use It
This one paid for itself immediately. We maintain 40+ n8n workflows and the node configuration gotchas are brutal — the Slack node v2.2 needs a specific parameter structure with messageType, the AI Agent node requires promptType:"define" when using the text parameter, the HTTP Request node has different behavior for neverError across versions. We had a growing list of these in our CLAUDE.md before n8n-MCP existed.
Now when Claude builds or modifies a workflow via the n8n API, it looks up the exact node properties instead of guessing. The validation feature is particularly useful — it catches things like an AI Agent node missing its memory connection or a ToolWorkflow node with the wrong schema configuration. Saved us from deploying broken workflows more than once.
n8nautomationworkflowslow-codeAI-agents