n8n MCP Integration

Connect to n8n's official MCP server to programmatically build, execute, and manage workflows.

Version Support

• n8n version: 2.16.1+
• MCP protocol: 2024-11-05
• Server name: n8n MCP Server v1.1.0

Configuration

MCP Server Setup

In n8n UI:

1. Go to Settings → n8n API
2. Create an API key for REST API access
3. Create an MCP token for MCP server access

Connection Config

{
  "mcpServers": {
    "n8n-mcp": {
      "type": "http",
      "url": "http://localhost:5678/mcp-server/http",
      "headers": {
        "Authorization": "Bearer <MCP_TOKEN>"
      }
    }
  }
}

Environment Variables

export N8N_MCP_URL="http://localhost:5678/mcp-server/http"
export N8N_MCP_TOKEN="<your-mcp-token>"

Available Tools (25)

Workflow Management

Workflow Execution

Workflow Creation (SDK)

Data Tables

Projects & Folders

Tool Parameters Reference

search_workflows

{
  "query": "string (optional) - Filter by name or description",
  "projectId": "string (optional) - Filter by project ID",
  "limit": "integer (optional, max 200) - Limit results"
}

execute_workflow

{
  "workflowId": "string (required) - The workflow ID to execute",
  "executionMode": "string (optional) - 'manual' for current version, 'production' for published version",
  "inputs": "object (optional) - Inputs for chat/form/webhook workflows"
}

get_execution

{
  "workflowId": "string (required)",
  "executionId": "string (required)",
  "includeData": "boolean (optional) - Include node execution data",
  "nodeNames": "array (optional) - Filter data by node names",
  "truncateData": "integer (optional) - Limit data items per node"
}

test_workflow

{
  "workflowId": "string (required)",
  "pinData": "object (required) - Pin data for nodes. Keys are node names, values are arrays of items wrapped in 'json': [{\"json\": {\"id\": \"123\"}}]",
  "triggerNodeName": "string (optional) - For multi-trigger workflows"
}

create_workflow_from_code

{
  "code": "string (required) - Validated SDK workflow code",
  "name": "string (optional) - Workflow name",
  "description": "string (optional, max 255 chars)",
  "projectId": "string (optional) - Project ID",
  "folderId": "string (optional) - Folder ID (requires projectId)"
}

search_nodes

{
  "queries": "array (required) - Search queries: service names, trigger types, or utility nodes"
}

get_node_types

{
  "nodeIds": "array (required) - Node IDs or objects with discriminators: [{\"nodeId\": \"...\", \"resource\": \"...\", \"operation\": \"...\"}]"
}

get_suggested_nodes

{
  "categories": "array (required) - Categories: chatbot, notification, scheduling, data_transformation, data_persistence, data_extraction, document_processing, form_input, content_generation, triage, scraping_and_research"
}

get_sdk_reference

{
  "section": "string (optional) - 'patterns', 'expressions', 'functions', 'rules', 'import', 'guidelines', 'design', or 'all' (default)"
}

create_data_table

{
  "projectId": "string (required)",
  "name": "string (required) - Unique table name",
  "columns": "array (required) - Column definitions: [{\"name\": \"...\", \"type\": \"...\"}]"
}

add_data_table_rows

{
  "dataTableId": "string (required)",
  "projectId": "string (required)",
  "rows": "array (required, max 1000) - Row objects: [{\"column1\": \"value1\", \"column2\": \"value2\"}]"
}

Usage Patterns

1. Creating a Workflow (Recommended Flow)

# Step 1: Get SDK reference
curl -X POST -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -H "Accept: application/json, text/event-stream" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_sdk_reference","arguments":{"section":"all"}}}' \
  "$N8N_MCP_URL"

# Step 2: Get suggested nodes for your use case
curl -X POST ... -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_suggested_nodes","arguments":{"categories":["notification","scheduling"]}}}'

# Step 3: Search specific nodes
curl -X POST ... -d '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"search_nodes","arguments":{"queries":["schedule trigger","slack","set"]}}}'

# Step 4: Get node types (CRITICAL - don't guess parameter names)
curl -X POST ... -d '{"jsonrpc":"2.0","id":4,"method":"tools/call","params":{"name":"get_node_types","arguments":{"nodeIds":["n8n-nodes-base.scheduleTrigger","n8n-nodes-base.slack"]}}}'

# Step 5: Validate workflow code
curl -X POST ... -d '{"jsonrpc":"2.0","id":5,"method":"tools/call","params":{"name":"validate_workflow","arguments":{"code":"..."}}}'

# Step 6: Create workflow
curl -X POST ... -d '{"jsonrpc":"2.0","id":6,"method":"tools/call","params":{"name":"create_workflow_from_code","arguments":{"code":"...","name":"My Workflow","description":"..."}}}'

2. Executing and Monitoring a Workflow

# Execute workflow
curl -X POST ... -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"execute_workflow","arguments":{"workflowId":"xxx"}}}'

# Get execution result (with data)
curl -X POST ... -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"get_execution","arguments":{"workflowId":"xxx","executionId":"yyy","includeData":true}}}'

3. Testing a Workflow (Before Publishing)

# Prepare test pin data
curl -X POST ... -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"prepare_test_pin_data","arguments":{"workflowId":"xxx"}}}'

# Test with pin data (pinData format is CRITICAL)
curl -X POST ... -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"test_workflow","arguments":{"workflowId":"xxx","pinData":{"NodeName":[{"json":{"field":"value"}}]}}}}'

4. Working with Data Tables

# Search data tables
curl -X POST ... -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"search_data_tables","arguments":{"query":"users"}}}'

# Add rows to data table
curl -X POST ... -d '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"add_data_table_rows","arguments":{"dataTableId":"xxx","projectId":"yyy","rows":[{"name":"John","email":"john@example.com"}]}}}'

SDK Workflow Example

import { workflow, trigger, node } from 'n8n-workflow-sdk';

export default workflow({
  name: 'Daily Slack Notification',
  description: 'Send daily summary to Slack',
  nodes: [
    trigger.schedule({
      name: 'Schedule',
      rule: { interval: [{ field: 'hours', hoursInterval: 24 }] }
    }),
    node.set({
      name: 'Prepare Message',
      values: { text: 'Daily report ready!' }
    }),
    node.slack({
      name: 'Send to Slack',
      resource: 'message',
      operation: 'send',
      channel: '#general',
      text: '={{ $node["Prepare Message"].json.text }}'
    })
  ],
  connections: [
    { from: 'Schedule', to: 'Prepare Message' },
    { from: 'Prepare Message', to: 'Send to Slack' }
  ]
});

MCP Protocol Notes

• Transport: HTTP with SSE (Server-Sent Events)
• Content-Type: application/json
• Accept: application/json, text/event-stream (required)
• Auth: Bearer token in Authorization header

Common Errors

Best Practices

1. Always get SDK reference first - Call get_sdk_reference before building workflows
2. Use get_node_types - Never guess parameter names; get exact types from the API
3. Validate before creating - Always call validate_workflow before create_workflow_from_code
4. Test before publishing - Use test_workflow with pin data to verify workflow logic
5. Use suggested nodes - Call get_suggested_nodes for curated recommendations by category
6. Correct pinData format - Items must be wrapped in json property: [{"json": {...}}]

References

• SDK Reference: See references/sdk-patterns.md for detailed SDK patterns
• Node Types: See references/common-nodes.md for frequently used nodes