What is N8n Mcp Agent Toolkit

n8n-mcp-agent-toolkit is a toolkit designed to enhance AI agent workflow support in n8n using the Model Context Protocol (MCP). It provides utilities, components, and patterns to facilitate the building and configuration of AI agents within n8n workflows.

Use cases

Use cases include automating data processing pipelines, enhancing customer support systems with AI chat capabilities, and integrating AI tools into business workflows for improved efficiency.

How to use

To use n8n-mcp-agent-toolkit, install it via npm with the command ‘npm install n8n-mcp-agent-toolkit’. You can then define AI node capabilities, create and format prompt templates, register tools, and handle data conversion within your workflows.

Key features

Key features include standardized AI node configuration, reusable prompt templates, unified tool configuration interfaces, and utilities for data handling between AI and non-AI nodes.

Where to use

n8n-mcp-agent-toolkit can be used in various fields that require AI integration into workflows, such as automation, data processing, and any application that benefits from AI-driven decision-making.

n8n-mcp-agent-toolkit

A toolkit for enhancing AI agent workflow support in n8n with the Model Context Protocol (MCP).

Overview

This toolkit provides utilities, components, and patterns to make it easier to build and configure AI agents within n8n workflows using the Model Context Protocol (MCP). It addresses several key areas:

1. AI Node Configuration: Standardized templates, validation, and error handling for AI nodes
2. Prompt Engineering: Reusable templates and schema definitions for prompt management
3. Tool Configuration: Unified interfaces for defining and managing tools for AI agents
4. Data Handling: Utilities for converting, validating, and formatting data between AI and non-AI nodes

Installation

npm install n8n-mcp-agent-toolkit

Usage

AI Node Configuration

Define node types and capabilities:

import { AiNodeCapability, isLanguageModelCapable } from 'n8n-mcp-agent-toolkit';

// Define node capabilities
const capabilities = [AiNodeCapability.LANGUAGE_MODEL, AiNodeCapability.CHAT_MODEL];

// Check if node has language model capability
if (isLanguageModelCapable(capabilities)) {
  // Configure language model options
}

Prompt Templates

Create and format prompt templates:

import { 
  WORKFLOW_AUTOMATION_EXPERT_PROMPT, 
  formatPromptTemplate 
} from 'n8n-mcp-agent-toolkit';

// Format a system prompt with variables
const formattedPrompt = formatPromptTemplate(
  WORKFLOW_AUTOMATION_EXPERT_PROMPT,
  {
    currentDate: new Date().toISOString(),
    workflowName: 'Data Processing Pipeline'
  }
);

Tool Configuration

Register and use tools:

import { 
  registerTool, 
  toolToJsonSchema,
  httpRequestTool
} from 'n8n-mcp-agent-toolkit';

// Register a built-in tool
registerTool(httpRequestTool);

// Convert to JSON Schema for LangChain or other frameworks
const schema = toolToJsonSchema(httpRequestTool);

Data Handling

Handle data conversion and validation:

import { 
  jsonToFlatRecord,
  validateWithZod,
  formatValue
} from 'n8n-mcp-agent-toolkit';

// Convert nested JSON to flat key-value pairs
const flatData = jsonToFlatRecord(complexObject);

// Validate data against a schema
const validation = validateWithZod(data, mySchema);

// Format values for display
const formattedDate = formatValue(new Date(), { dateFormat: 'long', includeTime: true });

Components

AI Node Configuration

• nodeTypes: Define and check AI node capabilities
• modelValidation: Validate AI model parameters
• errorHandling: Standardized error handling for AI nodes
• nodePropertyTemplates: Reusable n8n node properties

Prompt Templates

• schemas: Zod schemas for prompt templates
• systemPrompts: Pre-defined system prompts
• formatters: Format templates with variables
• chatPrompts: Chat-specific prompt templates

Tool Configuration

• toolSchema: Schema definitions for AI tools
• toolRegistry: Registry for managing tools
• builtInTools: Pre-defined tools

Data Handling

• adapters: Convert between data formats
• validators: Validate data
• formatters: Format data for display

Contributing

Contributions are welcome! Please check out our contributing guidelines.

License

MIT