n8n API MCP Server
About
MCP Server for interacting with an n8n instance
Details
- Author
- jasondsmith72
- GitHub stars
- 1
- Downloads
- 376
- Categories
- Developer Tools, API
Jump to
- Two local SQLite databases: API spec cache and Fast Memory.
- Search endpoints by path, summary, description, or tags.
- Execute API calls with automatic Fast Memory prompts.
- Natural language search that checks memory first, then spec.
- Manual save, list, delete, and clear Fast Memory entries.
- Raw request string execution (send_raw_api_request).
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
n8n API MCP ServerCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
Clone the repository, install dependencies (npm install), build the server (npm run build), and configure environment variables (N8N_URL, N8N_API_KEY) in your MCP client’s settings (e.g., cline_mcp_settings.json). Then load the API spec using load_api_spec_from_json before using search tools.
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"n8n api mcp server": {
"n8n-api-mcp": {
"command": "node",
"args": [
"C:\\MCP-Servers\\N8N-api-MCP\\build\\index.js"
],
"env": {
"N8N_URL": "",
"N8N_API_KEY": ""
}
}
}
}
}McpServers
{
"n8n-api-mcp": {
"command": "node",
"args": [
"C:\\MCP-Servers\\N8N-api-MCP\\build\\index.js"
],
"env": {
"N8N_URL": "",
"N8N_API_KEY": ""
}
}
}n8n API MCP Server
A Model Context Protocol (MCP) server designed to interact with a self-hosted n8n instance.
This server provides tools to:
- Execute calls to the n8n API.
- Search for n8n API endpoints using a local database populated from an OpenAPI specification.
- Store and retrieve frequently used or successful API calls using a "Fast Memory" database for quicker natural language access.
Features
Databases
- API Specification Database (api_spec.db): Stores details about n8n API endpoints, loaded from an OpenAPI/Swagger JSON file. Used by search and detail tools.
- Fast Memory Database (fast_memory.db): Caches successful API calls linked to natural language queries. Allows for quick retrieval of previously used calls and suggests saving new successful calls.
Tools
-search_api_endpoints: Search the local API spec database for endpoints matching a query (path, summary, description, tags).
- get_api_endpoint_details: Retrieve detailed information (parameters, request body, responses) for a specific endpoint (path and method) from the local API spec database.
- execute_api_call: Execute an API call to the configured n8n instance. Checks Fast Memory first; if not found, prompts to save the call to Fast Memory upon success.
- natural_language_api_search: Search for API calls using natural language. Checks Fast Memory first, then falls back to searching the API spec database based on the query.
- save_to_fast_memory: Manually save an API call (path, method, params, data) along with its original natural language query and a description to the Fast Memory database.
- list_fast_memory: List entries currently stored in the Fast Memory database, optionally filtering by a search term.
- delete_from_fast_memory: Delete a specific entry from the Fast Memory database by its ID.
- clear_fast_memory: Remove all entries from the Fast Memory database.
- load_api_spec_from_json: Load or update the API Specification database (api_spec.db) by parsing an OpenAPI/Swagger JSON file.
- send_raw_api_request: Execute an API call using a raw request string (e.g., "GET /workflows?limit=5"). Uses the same underlying logic as execute_api_call, including Fast Memory checks and save prompts.
Setup
1. Clone the repository:
git clone https://github.com/jasondsmith72/N8N-api-MCP.git
cd N8N-api-MCP
2. Install dependencies:
npm install
3. Build the server:
npm run build
This compiles the TypeScript code to JavaScript in the
build directory.
Configuration
This server requires environment variables to connect to your n8n instance. These are typically set in your MCP client's configuration file (e.g., cline_mcp_settings.json).
- N8N_URL: The base URL of your n8n instance (e.g., http://your-n8n.example.com). Defaults to http://localhost:5678 if not set.
- N8N_API_KEY: Your n8n API key for authentication.
Loading the API Specification
Before using the search tools (search_api_endpoints, natural_language_api_search), you need to populate the API Specification database.
1. Obtain the OpenAPI (Swagger) JSON specification file for your n8n version. This might be available from the n8n documentation or potentially downloadable from your instance (e.g., /api/v1/docs-json).
2. Use the load_api_spec_from_json tool, providing the absolute path to the downloaded JSON file:
load_api_spec_from_json(json_file_path="C:\\path\\to\\your\\n8n-openapi.json")
Installation for Cline / Claude Desktop
Add the following configuration to your cline_mcp_settings.json (adjust the path to index.js based on your clone location):
{
"mcpServers": {
"... other servers ...": {},
"n8n-api-mcp": {
"command": "node",
"args": [
"C:\\MCP-Servers\\N8N-api-MCP\\build\\index.js" // <-- Update this path if needed
],
"env": {
"N8N_URL": "YOUR_N8N_INSTANCE_URL", // e.g., "http://192.168.1.100:5678"
"N8N_API_KEY": "YOUR_N8N_API_KEY_HERE"
},
"disabled": false,
"autoApprove": []
}
}
}
YOUR_N8N_INSTANCE_URL and YOUR_N8N_API_KEY_HERE with your actual values)
Development
For development with auto-rebuild on file changes:
npm run watch
Debugging
Use the MCP Inspector for easier debugging:
npm run inspector
The Inspector provides a web interface to view MCP communication.
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
