n8n Comprehensive

Available MCP Tools

Workflow Management

• n8n_create_workflow - Create new workflows programmatically
• n8n_get_workflow - Retrieve workflow details and configuration
• n8n_update_partial_workflow - Update workflows incrementally
• n8n_update_full_workflow - Replace entire workflow configuration
• n8n_delete_workflow - Remove workflows
• n8n_list_workflows - List all workflows with metadata
• n8n_validate_workflow - Validate workflow structure
• n8n_autofix_workflow - Automatically fix common workflow errors

Node Discovery

• search_nodes - Find available node types by keyword
• get_node - Get detailed node configuration and parameters

Execution & Testing

• n8n_test_workflow - Test/trigger workflow execution
• n8n_executions - Manage workflow executions (get, list, delete)

Templates & Deployment

• search_templates - Search n8n template library
• get_template - Get template details
• n8n_deploy_template - Deploy templates directly to n8n instance

When to Use

• Building n8n workflows from scratch
• Implementing AI agents and chatbots with LangChain nodes
• Writing Code nodes for data transformation
• Integrating services and APIs into workflows
• Debugging workflow execution errors
• Using n8n API for programmatic control
• Working with vector stores and RAG implementations

Decision Trees

Data Transformation

• Simple field mapping → Expressions
• Complex logic → Code node
• Multiple transformations → Set node + Code node combo
• Batch processing → Loop + Code node

AI Implementation

• Simple text generation → Basic LLM Chain
• Question answering → Retrieval QA Chain
• Conversational AI → AI Agent + Memory
• Multi-step reasoning → ReAct Agent
• Database queries → SQL Agent

Integration Method

• Built-in node exists → Use built-in node
• Community node available → Install community node
• No node available → HTTP Request node
• Need custom logic → HTTP Request + Code node

Common Workflow Patterns

Pattern 1: Webhook to Database

Webhook Trigger → Validate Data (Code) → Transform (Set) → Insert to DB → Return Response

Pattern 2: Scheduled Data Sync

Schedule Trigger → Get Data from Source → Transform → Upsert to Destination → Log Results

Pattern 3: AI Document Processing

Trigger → Load Document → Extract Text → Chunk → Embed → Store in Vector DB

Pattern 4: Multi-Service Orchestration

Trigger → [Service A, B, C in parallel] → Merge Results → Transform → Output

Quick Reference: Expression Syntax

// Current item
{{ $json.fieldName }}
{{ $json.nested.property }}
{{ $json.array[0] }}

// Other nodes
{{ $('Node Name').item.json.field }}

// Dates
{{ $now }}
{{ $today.format('YYYY-MM-DD') }}

// Transformations
{{ $json.text.toUpperCase() }}
{{ $json.items.map(i => i.name) }}
{{ $json.price * 1.1 }}

Quick Reference: Code Node

// Basic structure
const items = $input.all();
return items.map(item => ({
  json: {
    // transformed data
  }
}));

// Access other nodes
const data = $('Previous Node').all();

// Binary preservation
return [{
  json: { processed: true },
  binary: item.binary
}];

Best Practices

• Workflow Design: Start simple, iterate to complex; use meaningful node names
• Code & Expressions: Prefer expressions for simple tasks; always return proper format
• AI Implementation: Choose appropriate model for task; use structured output parsers
• Integrations: Validate credentials first; handle rate limits gracefully
• Performance: Use batch processing; filter data early; monitor execution times