n8n Workflow Vectorization and Knowledge Graph

This project provides tools for analyzing, categorizing, and storing n8n workflows across multiple database types:

1. Vector Database (QDRANT): For semantic search and similarity matching
2. Graph Database (Neo4j): For relationship mapping and ontological representation
3. Relational Database (Supabase): For structured metadata storage

MCP Integrations

This project now includes Model Context Protocol (MCP) integrations for:

1. n8n: Import and manage workflows directly with n8n's API
2. QDRANT: Vector search and similarity matching for workflows
3. Supabase: Structured data storage for workflow metadata
4. Workflow Validation: Validate workflows against best practices

Project Structure

• workflow-parser.js: Processes n8n workflow JSON files, extracts metadata, and generates descriptions
• vectorize-workflows.js: Demonstrates how to use the QDRANT MCP server to vectorize workflows
• n8n-mcp-client.js: Client for interacting with the n8n MCP server
• n8n-workflow-validator.js: Validator for n8n workflows against best practices
• validate-workflow.js: CLI tool for validating workflows
• import-to-n8n.js: Tool for importing workflows to n8n
• processed-workflows/: Contains the processed workflow data with enriched metadata
• workflows/: Directory containing original n8n workflow files
• MCP Servers located in /Users/kinglerbercy/MCP/:

Setup and Installation

1. Process Workflows

The workflow-parser.js script processes n8n workflow files, extracting metadata and generating descriptions:

npm run parse
# or
node workflow-parser.js

This creates enriched workflow data in the processed-workflows/ directory.

2. Configure Environment Variables

Copy the example environment file and update with your settings:

cp .env.example .env
# Edit .env with your configuration

3. MCP Integrations

QDRANT MCP Integration

# Connect to QDRANT for vectorization and semantic search
npm run vectorize-mcp

Supabase MCP Integration

# Store workflows in Supabase
npm run supabase-mcp

n8n MCP Integration

# Connect to n8n and list workflows
npm run n8n-demo

# Import all workflows to n8n
npm run import-all

# Import a specific workflow to n8n
npm run import <workflow-id>
# or
npm run import <workflow-file.json>

Workflow Validation

# Validate a workflow in n8n by ID
npm run validate <workflow-id>

# Validate a local workflow file
npm run validate -f <workflow-file.json>

# Validate with high strictness
npm run validate -s high <workflow-id>

# Run specific validators only
npm run validate -v naming,security,errorHandling <workflow-id>

The workflow validator checks for best practices in five categories:

1. Naming: Ensures workflows and nodes have descriptive names
2. Error Handling: Checks for proper error handling nodes and patterns
3. Security: Validates credential usage and security practices
4. Performance: Identifies potential performance issues
5. Documentation: Ensures workflows are properly documented

4. Manual Vectorization (Legacy Method)

The QDRANT MCP server can also be configured in the Cline MCP settings:

"n8n-workflows-qdrant": {
  "command": "node",
  "args": [
    "/Users/kinglerbercy/MCP/qdrant-server/index.js"
  ],
  "env": {
    "OPENAI_API_KEY": "your-openai-api-key",
    "QDRANT_URL": "http://localhost:6333"
  },
  "disabled": false,
  "autoApprove": []
}

After restarting Cline with the MCP server configuration, you can use the MCP tools to vectorize and search workflows:

// Vectorize workflows
use_mcp_tool(
  server_name: "n8n-workflows-qdrant",
  tool_name: "vectorize_workflows",
  arguments: {
    "workflowsDir": "/path/to/workflows",
    "processedDir": "/path/to/processed-workflows"
  }
)

// Search for similar workflows
use_mcp_tool(
  server_name: "n8n-workflows-qdrant",
  tool_name: "search_similar_workflows",
  arguments: {
    "query": "workflow that handles email automation",
    "limit": 5,
    "filter": {
      "category": "workflow",
      "tags": ["email", "automation"],
      "complexity": "moderate"
    }
  }
)

Workflow Analysis

The workflow parser extracts the following information:

1. Category: Derived from the filename prefix (e.g., "agent:", "workflow:", "tool:")
2. Name: Extracted from the filename
3. Description: Generated from workflow content, including node types and documentation
4. Tags: Extracted from node types and services used
5. Complexity: Analyzed based on node count, connection count, and unique node types

Database Integration

Vector Database (QDRANT)

• Purpose: Semantic search and similarity matching
• Implementation: QDRANT MCP server with OpenAI embeddings
• Features: Search by query, filter by category, tags, and complexity

Graph Database (Neo4j)

• Purpose: Relationship mapping and ontological representation
• Implementation: To be implemented
• Features: Query relationships between workflows, node types, and categories

Relational Database (Supabase)

• Purpose: Structured metadata storage
• Implementation: To be implemented
• Features: SQL queries for structured data analysis

Available MCP Tools

QDRANT MCP Tools

The QDRANT MCP server provides the following tools:

1. vectorize_workflows: Vectorize n8n workflows and store them in Qdrant
2. search_similar_workflows: Search for similar workflows based on a query
3. get_workflow_categories: Get all workflow categories
4. get_workflow_tags: Get all workflow tags

n8n MCP Tools

The n8n MCP server provides the following tools:

1. list_workflows: Get all workflows from n8n
2. get_workflow: Get a specific workflow by ID
3. create_workflow: Create a new workflow
4. update_workflow: Update an existing workflow
5. delete_workflow: Delete a workflow
6. import_workflow: Import a workflow from a file
7. export_workflow: Export a workflow to a file

Workflow Validation Tools

The workflow validator provides the following validation categories:

1. naming: Validate node and workflow naming conventions
2. errorHandling: Validate error handling patterns
3. security: Validate security practices and credential usage
4. performance: Identify potential performance issues
5. documentation: Ensure proper workflow documentation

Future Enhancements

1. Implement Neo4j integration for relationship mapping
2. Implement Supabase integration for structured metadata storage
3. Create a web interface for browsing and searching workflows
4. Add more advanced analytics and visualization tools