A comprehensive tutorial series for building AI agents with n8n, progressing from simple chatbots to advanced multi-agent systems.
This tutorial covers:
- Basic Agent: Simple conversational AI agent
- API Integration: Agent that interacts with external services
- Memory-Enabled Agent: Stateful agent that remembers conversation context
- MCP Integration: Agent using Model Context Protocol for tool use
- Multi-Agent Orchestration: Coordinated system with specialized agents
- n8n installed (local or cloud)
- Google Gemini API key (FREE tier available!) - Get key
- Basic understanding of n8n workflows (see n8n basics guide)
💡 New to n8n? Check the n8n Basics Guide - we only assume you can click, drag, and copy/paste!
✅ 100% FREE: All tutorials use Google Gemini API with a generous free tier! See docs/setup.md for setup instructions.
ai-agent101/
├── README.md # This file
├── docs/ # Documentation
│ ├── setup.md # Setup instructions
│ ├── concepts.md # Core concepts
│ └── troubleshooting.md # Common issues
├── 01-basic-agent/ # Level 1: Basic agent
│ ├── README.md
│ └── workflow.json
├── 02-api-integration/ # Level 2: API integration
│ ├── README.md
│ └── workflow.json
├── 03-memory-agent/ # Level 3: Stateful agent
│ ├── README.md
│ └── workflow.json
├── 04-mcp-agent/ # Level 4: MCP integration
│ ├── README.md
│ └── workflow.json
└── 05-orchestration/ # Level 5: Multi-agent system
├── README.md
└── workflow.json
-
Setup n8n
# Using npx (recommended for beginners) npx n8n # Or using Docker docker run -it --rm --name n8n -p 5678:5678 n8nio/n8n
-
Configure Credentials
- Add your Google Gemini API key in n8n credentials
- Set up any additional services needed (Redis for Tutorial 03)
-
Import Workflows
- Navigate to each tutorial folder
- Import the
workflow.jsonfile into n8n - Workflows come with Google Gemini pre-configured! Just add your credential.
- Follow the README in each folder
✅ Ready to Use: All workflows are pre-configured with FREE Google Gemini! Just import and add your API key.
Difficulty: Beginner Time: 15 minutes
Learn to create a simple conversational AI agent that:
- Receives user input via webhook
- Processes queries with an LLM
- Returns responses
Key Concepts: Webhooks, LLM nodes, basic prompting
Difficulty: Beginner-Intermediate Time: 30 minutes
Build an agent that interacts with external APIs:
- Weather information lookup
- Function calling/tool use
- Response formatting
Key Concepts: HTTP requests, function calling, data transformation
Difficulty: Intermediate Time: 45 minutes
Create a stateful agent with conversation memory:
- Session management
- Context retention across messages
- Redis/database integration
Key Concepts: State management, session storage, context windows
Difficulty: Intermediate-Advanced Time: 60 minutes
Implement an agent using Model Context Protocol:
- MCP server integration
- Tool discovery and execution
- Standardized context sharing
Key Concepts: MCP protocol, tool servers, context management
Difficulty: Advanced Time: 90 minutes
Build a coordinated multi-agent system:
- Specialized agents (research, writer, reviewer)
- Agent coordination and handoffs
- Complex workflow orchestration
Key Concepts: Agent roles, task delegation, workflow coordination
We recommend following the tutorials in order:
Basic Agent → API Integration → Memory → MCP → Orchestration
↓ ↓ ↓ ↓ ↓
Chat Functions State Tools Teamwork
Found an issue or want to improve the tutorials? Contributions welcome!
MIT License - Feel free to use these examples for learning and building your own agents.
Ready to start? Head to 01-basic-agent to begin your journey!