n8n Workflow Builder MCP Server

Build n8n workflows from text using AI - Built by FlowEngine

Turn natural language into production-ready, validated n8n workflows. This isn't just another n8n context provider, it's a complete workflow generation engine with built-in validation, auto-fixing, and architectural intelligence.

this MCP is built to provide validated n8n workflows — not just context.

13-Point Validation Engine - Catches errors before you import
Auto-Fix Malformed Workflows — Automatically repairs common issues
Architecture Recommendations — Suggests optimal workflow patterns
Security Scanning — Detects credential leaks and vulnerabilities
Performance Analysis — Identifies bottlenecks and optimization opportunities
Real Parameter Schemas — Loaded directly from n8n packages for accuracy
600+ Registered Node Types — Only real nodes, no hallucinations

Result: Workflows that actually import and run on first try.

Quick Start

Local Install | Remote Server (no install)

Video Demo

Feature Details

13-Point Validation Engine

Every workflow passes through 13 validation checks before output:

Node Type Validation — Verifies every node type exists in n8n
Connection Integrity — Ensures all connections reference existing nodes
Parameter Type Checking — Validates parameter types match node schemas
Required Fields — Checks all required parameters are present
Credential References — Validates credential configurations
Expression Syntax — Checks n8n expression syntax (={{ }})
Position Validation — Ensures nodes have valid canvas positions
Duplicate Detection — Catches duplicate node names
Orphan Node Detection — Finds disconnected nodes
Trigger Validation — Ensures workflows have proper entry points
Loop Detection — Identifies potential infinite loops
Output Mapping — Validates data flow between nodes
Version Compatibility — Checks node version compatibility

Auto-Fix Malformed Workflows

When validation finds issues, the engine automatically repairs them:

Missing positions → Auto-calculates layout on canvas
Invalid node names → Generates unique, valid names
Missing connections array → Initializes proper structure
Incorrect typeVersion → Updates to current supported version
Malformed parameters → Applies sensible defaults
Broken JSON structure → Attempts recovery and repair

Architecture Recommendations

Based on your task description, the engine analyzes keywords and intent to suggest the optimal pattern. Here's when each is recommended:

Regular Workflows (Deterministic)

Linear Pipeline

When: Simple, predictable data transformations
Example: "Get data from API → Transform → Save to database"
Best for: ETL, data sync, scheduled reports
Why not AI: No decision-making needed, faster execution, lower cost

Conditional Branching

When: Known decision points with clear rules
Example: "If order > $100, send to manager; else auto-approve"
Best for: Approval flows, routing, rule-based automation
Why not AI: Rules are explicit, no reasoning required

Parallel Processing

When: Independent operations that can run simultaneously
Example: "Send email AND update CRM AND log to Slack"
Best for: Notifications, multi-system updates, batch processing
Why not AI: No dependencies between branches

Event-Driven

When: Reacting to external triggers
Example: "When webhook received → process → respond"
Best for: API endpoints, real-time integrations, chatbots
Why not AI: Response is formulaic, not conversational

AI Agent Workflows (Autonomous)

When to use AI Agents instead of regular workflows:

Use AI Agent When...	Use Regular Workflow When...
Task requires reasoning	Steps are predictable
Input is unstructured (natural language)	Input is structured (JSON, forms)
Multiple tools might be needed dynamically	Tool sequence is known
Conversation/context matters	Stateless processing
Decision logic is complex or fuzzy	Rules are explicit

Single AI Agent

When: One autonomous entity with access to tools
Example: "Customer support bot that can search docs, create tickets, and escalate"
Architecture: Chat Trigger → AI Agent (with tools) → Response
Tools: Calculator, Code, HTTP Request, custom tools
Memory: Optional conversation memory for context

AI Agent with Memory

When: Conversation context matters across messages
Example: "Personal assistant that remembers user preferences"
Architecture: Chat Trigger → Memory Load → AI Agent → Memory Save → Response
Memory types: Buffer (recent), Window (last N), Summary (compressed)

Multi-Agent Workflows

For complex tasks, multiple specialized agents can collaborate:

Sequential Multi-Agent

When: Task has distinct phases requiring different expertise
Example: "Research Agent → Analysis Agent → Writing Agent"
Flow: Agent 1 output becomes Agent 2 input
Best for: Content pipelines, multi-step reasoning

Supervisor Pattern

When: Need coordination between specialist agents
Example: "Supervisor routes to: Sales Agent, Support Agent, or Technical Agent"
Flow: Supervisor Agent decides which specialist handles the task
Best for: Customer service, complex routing

Parallel Agents

When: Multiple perspectives needed simultaneously
Example: "Analyst Agent AND Risk Agent AND Compliance Agent all review"
Flow: Split → Multiple Agents → Merge results
Best for: Review processes, multi-criteria evaluation

Hierarchical Agents

When: Complex orchestration with sub-tasks
Example: "Manager Agent delegates to Team Agents who use Tool Agents"
Flow: Top-level agent breaks down task, delegates, aggregates
Best for: Large-scale automation, enterprise workflows

How the Engine Decides

The recommendation engine analyzes your description for:

AI keywords: "chat", "conversation", "understand", "decide", "reason" → suggests AI Agent
Automation keywords: "sync", "transform", "schedule", "trigger" → suggests Regular Workflow
Complexity signals: Multiple conditions, dynamic routing → suggests Conditional or Agent
Tool mentions: "search", "calculate", "call API" → adds appropriate tools to Agent
Memory signals: "remember", "context", "history" → adds memory to Agent

Security Scanning

Detects vulnerabilities before deployment:

Hardcoded credentials — API keys, passwords, tokens in plain text
Insecure protocols — HTTP instead of HTTPS
Missing authentication — API calls without auth headers
Sensitive data exposure — PII in logs or outputs
Code injection risks — Unsafe code node patterns

Performance Analysis

Identifies bottlenecks and optimization opportunities:

Execution time estimates — Per-node and total workflow
Parallel path detection — Opportunities for concurrent execution
API call optimization — Batch vs. individual requests
Memory usage patterns — Large data handling concerns
Rate limiting risks — High-frequency API calls

Real Parameter Schemas

Parameters are loaded directly from n8n packages, not guessed:

Accurate defaults — Real default values from node definitions
Correct types — String, number, boolean, options validated
Required vs optional — Knows which fields are mandatory
Nested structures — Complex parameter objects handled correctly

600+ Registered Node Types

No hallucinations. The AI can only use nodes that actually exist:

Every node type is loaded from the official n8n-nodes-base package
LangChain nodes from @n8n/n8n-nodes-langchain included
Node registry is generated directly from n8n source code
If a node doesn't exist in the registry, it won't be used
Prevents the common AI problem of inventing fake node types

Quick Start

Installation Methods

Choose the method that works best for you:

1. Smithery (One-Click Install)

Easy local installation for Claude Desktop. View on Smithery

npx @smithery/cli install @Ami3466/mcp-flowengine-n8n-workflow-builder --client claude

Note: Smithery has a 50 uses/day limit. For unlimited access, use the remote server below.

2. Remote Server (More Stable - Unlimited) ⭐

Connect to our hosted server - no installation required!

Add this to your Claude Desktop config:

Mac: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "flowengine-n8n": {
      "url": "https://mcp-flowengine-n8n-workflow-builder.onrender.com/mcp",
      "transport": "http"
    }
  }
}

Benefits:

✅ More stable - Dedicated hosting infrastructure
✅ Unlimited usage - No daily limits
✅ Always up-to-date - Automatically updated
✅ Zero installation - Just add config and restart Claude
✅ Free - Community hosted

3. npm (Manual Install)

For all other MCP clients:

npm install -g flowengine-n8n-workflow-builder

Setup by Platform

Claude Desktop

Option A: Smithery (Local Install)

View on Smithery

npx @smithery/cli install @Ami3466/mcp-flowengine-n8n-workflow-builder --client claude

Restart Claude Desktop after installation.

Note: Limited to 50 uses/day. For unlimited access, use the remote server below.

Option B: Remote Server (More Stable - Unlimited) ⭐

No installation required! Just add config and restart.

Edit Claude Desktop config:
- Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%/Claude/claude_desktop_config.json

Add this configuration:

{
  "mcpServers": {
    "flowengine-n8n": {
      "url": "https://mcp-flowengine-n8n-workflow-builder.onrender.com/mcp",
      "transport": "http"
    }
  }
}

Restart Claude Desktop (fully quit and reopen)
Verify connection:
- Look for the 🔌 MCP icon in Claude Desktop
- You should see "flowengine-n8n" listed with green status

Benefits: More stable, unlimited usage, no local installation, always updated.

Option C: Manual Local Install

Install the package:

npm install -g flowengine-n8n-workflow-builder

Edit Claude Desktop config:
- Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
- Windows: %APPDATA%/Claude/claude_desktop_config.json

Add this configuration:

{
  "mcpServers": {
    "flowengine-n8n": {
      "command": "flowengine-n8n"
    }
  }
}

Restart Claude Desktop (fully quit and reopen)

Claude Code (VS Code Extension)

Install the package globally:

npm install -g flowengine-n8n-workflow-builder

Open VS Code Settings (Cmd/Ctrl + ,)
Search for "MCP"

Add MCP Server:

Click "Edit in settings.json"
Add to claude.mcpServers:

{
  "claude.mcpServers": {
    "flowengine-n8n": {
      "command": "flowengine-n8n"
    }
  }
}

Reload VS Code (Cmd/Ctrl + Shift + P → "Developer: Reload Window")
Start using:
- Open Claude Code panel
- Ask Claude to build n8n workflows
- The MCP server will be automatically available

Cursor

Install the package:

npm install -g flowengine-n8n-workflow-builder

Open Cursor Settings:
- Mac: Cursor → Settings → Features
- Windows/Linux: File → Preferences → Features
Navigate to MCP Settings:
- Scroll to "Model Context Protocol"
- Click "Edit Config"

Add configuration:

{
  "mcpServers": {
    "flowengine-n8n": {
      "command": "flowengine-n8n"
    }
  }
}

Restart Cursor
Verify:
- Open Cursor's AI chat
- The MCP server should be available
- Try: "Create an n8n workflow for me"

Cline (VS Code)

Install the package:

npm install -g flowengine-n8n-workflow-builder

Open Cline Settings in VS Code:
- Open Command Palette (Cmd/Ctrl + Shift + P)
- Type "Cline: Open Settings"

Add MCP Server:

In Cline settings, find "MCP Servers"
Add new server:

{
  "flowengine-n8n": {
    "command": "flowengine-n8n"
  }
}

Reload VS Code

Continue.dev

Install the package:

npm install -g flowengine-n8n-workflow-builder

Open Continue config:
- Open Command Palette (Cmd/Ctrl + Shift + P)
- Type "Continue: Open config.json"

Add MCP server to config:

{
  "mcpServers": {
    "flowengine-n8n": {
      "command": "flowengine-n8n"
    }
  }
}

Reload Continue extension

What This Does

🤖 Turns Text into Workflows

Describe what you want in plain language:

"Monitor my email and notify me on Slack"
"Build an AI chatbot with memory and tools"
"Sync data between Google Sheets and my database"
"Create a customer support automation workflow"

Your AI assistant generates complete, working n8n workflows.

🧠 Expert n8n Knowledge

Your AI gets access to:

600+ Node Types - All n8n-nodes-base and LangChain nodes
Real Parameter Schemas - Loaded directly from n8n packages
Best Practices - Workflow design patterns and optimizations
Intelligent Validation - Automatic error detection and fixing

✨ Powerful Features

Workflow Generation:

Build workflows from natural language descriptions
Add, edit, and delete nodes programmatically
Connect nodes and manage workflow structure
Get detailed workflow analysis

Intelligence & Suggestions:

Architecture recommendations (linear, conditional, AI agent, etc.)
Node suggestions for specific tasks
Workflow analysis and improvements
Natural language explanations

Quality & Security:

Comprehensive workflow validation
Security vulnerability scanning
Performance analysis and bottleneck detection
Dry-run workflow testing

Templates & Search:

Pre-built workflow templates
Search 600+ nodes by keyword
Browse nodes by category
Get detailed node documentation

Usage Examples

Once installed, ask your AI assistant to help with workflows:

Create a New Workflow

"Create an n8n workflow that monitors Gmail for emails with 'urgent' in the subject and sends a Slack notification to #alerts"

Analyze Existing Workflow

"Analyze this workflow and suggest improvements"
[paste your workflow JSON]

Get Node Recommendations

"What nodes should I use to build a customer onboarding automation?"

Validate and Fix

"Validate this workflow and fix any errors"
[paste workflow JSON]

Security Scan

"Scan this workflow for security issues"
[paste workflow JSON]

Available Tools

Your AI assistant gets access to 23 powerful tools:

Workflow Building:

build_workflow - Generate workflows from descriptions
add_node - Add nodes to existing workflows
edit_node - Modify node parameters
delete_node - Remove nodes
add_connection - Connect nodes
get_workflow_details - Analyze workflow structure

Validation & Quality:

validate_workflow - Comprehensive validation with auto-fix
test_workflow - Dry-run simulation
scan_security - Security vulnerability detection
analyze_performance - Performance and bottleneck analysis

Intelligence:

suggest_architecture - Recommend workflow patterns
suggest_nodes - Node recommendations for tasks
analyze_workflow - Deep workflow insights
suggest_improvements - Optimization suggestions
explain_workflow - Natural language explanations

Node Library:

search_nodes - Search 600+ nodes
list_nodes_by_category - Browse by category
get_node_details - Detailed node documentation

Templates:

list_templates - Browse workflow templates
get_template - Get specific templates
search_templates - Search templates

Utilities:

extract_workflow_json - Extract JSON from text
fix_json - Repair malformed JSON

Deploy Your Workflows

Option 1: FlowEngine.cloud (Recommended)

Build for free, deploy instantly:

Generate workflow using this MCP server
Visit flowengine.cloud
Import your workflow JSON
Test and deploy - no infrastructure needed

Why FlowEngine.cloud?

✅ No server setup or management
✅ Built-in monitoring and logs
✅ Automatic scaling
✅ Visual workflow editor
✅ Free tier available

Option 2: Self-Hosted n8n

Generate workflow using this MCP server
Open your n8n instance
Import JSON (... → Import from File)
Configure credentials and activate

Troubleshooting

MCP Server Not Showing Up?

Verify installation:

which flowengine-n8n
# Should show: /usr/local/bin/flowengine-n8n (or similar)

Test the server manually:

flowengine-n8n
# Should start the MCP server

Check Claude Desktop logs:
- Mac: ~/Library/Logs/Claude/mcp*.log
- Windows: %APPDATA%/Claude/logs/mcp*.log
Restart your AI client completely (fully quit and reopen)

Package Not Found?

# Update npm
npm install -g npm@latest

# Reinstall the package
npm uninstall -g flowengine-n8n-workflow-builder
npm install -g flowengine-n8n-workflow-builder

Permission Issues?

Mac/Linux:

sudo npm install -g flowengine-n8n-workflow-builder

Windows: Run PowerShell/CMD as Administrator

Smithery Installation Issues?

# Clear npm cache
npm cache clean --force

# Try manual installation instead
npm install -g flowengine-n8n-workflow-builder

How It Works

This MCP server connects your AI assistant to expert n8n knowledge:

You describe what you want
         ↓
Your AI Tool (Claude/Cursor/etc.)
         ↓
   MCP Protocol
         ↓
FlowEngine n8n Builder
         ↓
Expert n8n Knowledge + Validation
         ↓
Complete, Working Workflow

What Your AI Gets:

Deep n8n expertise
600+ node definitions with real schemas
Workflow patterns and best practices
Validation and auto-fixing capabilities
Security and performance analysis

What You Get:

Production-ready workflows
Properly validated JSON
Working configurations
Best practice implementations

Features

✅ Build workflows from text - Natural language to working n8n workflows ✅ 600+ Node Types - Full n8n-nodes-base and LangChain support ✅ Real Parameter Schemas - Loaded directly from n8n packages ✅ Intelligent validation - Automatic error detection and fixing ✅ Security scanning - Detect vulnerabilities and sensitive data ✅ Performance analysis - Find bottlenecks and optimize ✅ Works with any LLM - Universal MCP protocol support ✅ No API keys needed - Works completely offline ✅ Built by FlowEngine - Production-tested technology ✅ Deploy anywhere - FlowEngine.cloud or self-hosted

About FlowEngine

FlowEngine is a platform for building and deploying n8n workflows with AI:

🎨 Visual Builder - Drag-and-drop editor at flowengine.cloud
🤖 AI-Powered - Generate workflows with natural language
☁️ Managed Hosting - Deploy instantly, no DevOps needed
📈 Production Ready - Monitoring, logs, and scaling included
🆓 Free Tier - Try and build workflows for free

This MCP server brings FlowEngine's workflow generation technology to your local development environment.

Use Cases

For Developers:

Generate boilerplate workflows quickly
Prototype automation ideas fast
Learn n8n patterns and best practices
Build complex workflows with AI assistance

For Teams:

Accelerate workflow development
Standardize workflow patterns
Reduce learning curve for n8n
Scale automation initiatives

For Businesses:

Automate repetitive tasks
Connect systems and tools
Build custom integrations
Deploy AI-powered workflows

Support & Resources

FlowEngine Platform: flowengine.cloud
Documentation: docs.flowengine.cloud
npm Package: View on npm
Smithery Registry: Install via Smithery
GitHub Repository: View Source
Report Issues: Contact FlowEngine support

License

MIT License with Commons Clause

✅ You CAN:

Use for personal projects (free forever)
View and study the source code
Modify for your own personal use
Contribute improvements back

❌ You CANNOT:

Sell this software
Offer as a commercial/paid service
Build competing products using this code
Use in any commercial product or service

For commercial licensing: flowengine.cloud

Full license: LICENSE

Built by FlowEngine - Enterprise-grade n8n workflow automation platform

Website • Documentation • npm • Smithery

Name		Name	Last commit message	Last commit date
Latest commit History 16 Commits
prompts		prompts
render-deploy		render-deploy
scripts		scripts
src		src
.dockerignore		.dockerignore
.gitignore		.gitignore
.npmignore		.npmignore
CHANGELOG.md		CHANGELOG.md
DEPLOYMENT.md		DEPLOYMENT.md
Dockerfile		Dockerfile
LICENSE		LICENSE
PUBLISH.md		PUBLISH.md
README.md		README.md
package-lock.json		package-lock.json
package.json		package.json
render.yaml		render.yaml
smithery.yaml		smithery.yaml
test-tools.mjs		test-tools.mjs
tsconfig.json		tsconfig.json

License

FlowEngine-cloud/mcp-n8n-workflow-builder-flowengine

Folders and files

Latest commit

History

Repository files navigation

n8n Workflow Builder MCP Server

Quick Start

Video Demo

Feature Details

13-Point Validation Engine

Auto-Fix Malformed Workflows

Architecture Recommendations

Regular Workflows (Deterministic)

AI Agent Workflows (Autonomous)

How the Engine Decides

Security Scanning

Performance Analysis

Real Parameter Schemas

600+ Registered Node Types

Quick Start

Installation Methods

1. Smithery (One-Click Install)

2. Remote Server (More Stable - Unlimited) ⭐

3. npm (Manual Install)

Setup by Platform

Claude Desktop

Option A: Smithery (Local Install)

Option B: Remote Server (More Stable - Unlimited) ⭐

Option C: Manual Local Install

Claude Code (VS Code Extension)

Cursor

Cline (VS Code)

Continue.dev

What This Does

🤖 Turns Text into Workflows

🧠 Expert n8n Knowledge

✨ Powerful Features

Usage Examples

Create a New Workflow

Analyze Existing Workflow

Get Node Recommendations

Validate and Fix

Security Scan

Available Tools

Deploy Your Workflows

Option 1: FlowEngine.cloud (Recommended)

Option 2: Self-Hosted n8n

Troubleshooting

MCP Server Not Showing Up?

Package Not Found?

Permission Issues?

Smithery Installation Issues?

How It Works

Features

About FlowEngine

Use Cases

Support & Resources

License

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Contributors 2

Languages

Packages