Stop letting your assistant hallucinate n8n node parameters — this server hands it the real schemas, templates and validation.
n8n-MCP gives AI assistants indexed documentation for thousands of n8n nodes plus validation and, optionally, live workflow management against your own instance.
Node.js for the npx path, or Docker for container deployments. The documentation tools work with no credentials; workflow management additionally needs N8N_API_URL and an n8n API key.
Compare MCP servers by category, or browse hosted MCP servers if you prefer an endpoint over a local process.
The assistant searches templates first, falls back to node discovery, validates each node configuration, then emits a workflow JSON that actually imports cleanly.
Validate connections and expressions against the real node schemas to find the unset parameter or bad reference, rather than reading stack traces from n8n's execution log.
Run with a read-only API key and DISABLED_TOOLS set so an agent can list workflows, read executions and audit configuration without any ability to modify them.
Ask which nodes can act as AI tools, what a node's authentication properties are, or how a given operation is configured, without tab-hopping through docs.
Check the linked publisher documentation and your connected client's tool list. Names and availability can vary with the server version and account permissions.
| Tool | What it does |
|---|---|
| tools_documentation | Returns usage documentation and best practices for the server's own tools — the recommended first call. |
| search_nodes | Full-text search across the node catalogue, with source filters for community or verified nodes and an option to include real example configurations. |
| get_node | Unified node lookup with minimal, standard and full detail levels plus modes for property search, markdown docs, version comparison and breaking changes. |
| validate_node | Validate a node configuration, either as a fast required-fields check or a full pass using a chosen validation profile. |
| validate_workflow | Validate an entire workflow including connections, expressions and AI agent wiring. |
| search_templates | Find workflow templates by keyword, by node types used, by curated task, or by metadata such as complexity and target audience. |
| get_template | Retrieve a template's workflow JSON at nodes-only, structure or full fidelity. |
| n8n_create_workflow | Create a new workflow with nodes and connections on the connected instance. |
| n8n_get_workflow | Retrieve a workflow at full, details, structure or minimal fidelity. |
| n8n_update_partial_workflow | Apply batched diff operations to a workflow without resending the whole JSON. |
| n8n_update_full_workflow | Replace an entire workflow definition. |
| n8n_delete_workflow | Permanently delete a workflow. |
| n8n_list_workflows | List workflows with filtering and pagination. |
| n8n_validate_workflow | Validate a workflow that already lives on the instance, by id. |
| n8n_autofix_workflow | Automatically repair common structural errors in a deployed workflow. |
| n8n_workflow_versions | Inspect version history and roll a workflow back. |
| n8n_deploy_template | Deploy a template from n8n.io straight to your instance with autofix applied. |
| n8n_test_workflow | Trigger a workflow execution via webhook, form or chat entry point. |
| n8n_executions | List, fetch and delete workflow executions. |
| n8n_evaluations | Run and read evaluation test runs, including aggregated metrics and per-case results. |
| n8n_manage_datatable | List, get, create, update and delete n8n data tables and their rows. |
| n8n_manage_credentials | Manage credentials, including schema lookup. Frequently disabled outright because even reads expose sensitive material. |
| n8n_audit_instance | Security audit combining n8n's built-in audit API with a deeper workflow scan. |
| n8n_health_check | Check n8n API connectivity and report available features. |
{
"mcpServers": {
"n8n-mcp": {
"command": "npx",
"args": ["n8n-mcp"],
"env": {
"MCP_MODE": "stdio",
"LOG_LEVEL": "error",
"DISABLE_CONSOLE_OUTPUT": "true",
"N8N_API_URL": "https://your-n8n-instance.com",
"N8N_API_KEY": "your-api-key"
}
}
}
}Node.js for the npx path, or Docker for container deployments. The documentation tools work with no credentials; workflow management additionally needs N8N_API_URL and an n8n API key.
| Variable | Description | Required |
|---|---|---|
| MCP_MODE | Set to stdio for desktop clients so only JSON-RPC reaches stdout; without it clients report JSON parse errors.e.g. stdio | Yes |
| N8N_API_URL | Base URL of your n8n instance. Required to unlock the workflow management tools.e.g. https://your-n8n-instance.com | Optional |
| N8N_API_KEY | n8n API key created under Settings then API. Use a read-only key when you only want inspection. | Optional |
| N8N_CF_CLIENT_ID | Cloudflare Access service-token client ID, if your n8n instance sits behind Zero Trust. | Optional |
| N8N_CF_CLIENT_SECRET | Cloudflare Access service-token secret, sent only to the N8N_API_URL origin. | Optional |
| DISABLED_TOOLS | Comma-separated tool names to remove entirely — the recommended way to build a read-only deployment.e.g. n8n_delete_workflow,n8n_manage_credentials | Optional |
| DISABLED_TOOL_OPERATIONS | Blocks individual operations inside tools that mix reads and writes, e.g. n8n_executions:delete.e.g. n8n_executions:delete | Optional |
| LOG_LEVEL | Logging verbosity; error keeps stdio clean.e.g. error | Optional |
Read, search and edit files inside allowed directories with the Model Context Protocol reference Filesystem server.
Retrieve version-aware library documentation and code examples with Context7 MCP, using resolve-library-id and query-docs.
Food data tools for any AI agent. Describe food in English, get structured data back.
MCP server for instant URL indexing via IndexNow and Google Indexing API
# Supabase MCP Server <p align="center"> <picture> <source media="(prefers-color-scheme: dark)"
Make line-based edits to a text file. Each edit replaces exact line sequences.