n8n Tutorial for Beginners 2026: Build Your First AI Workflow

n8n is an open-source workflow automation platform that lets developers and technical teams build automated pipelines — including AI-powered ones — without writing code for every integration. This guide walks you from zero to a working AI workflow in about 30 minutes, covering setup, core concepts, and two hands-on builds you can run today.

What Is n8n? (The Open-Source AI Workflow Platform Built for Developers)

n8n is an open-source, self-hostable workflow automation platform designed for developers who need the flexibility of code without the overhead of building every integration from scratch. Unlike purely no-code tools like Zapier, n8n gives you a visual workflow editor plus direct access to JavaScript and Python in any node — so you control exactly what happens with your data. As of 2026, n8n 2.0 ships with native LangChain integration and 70+ AI nodes, making it a first-class platform for building AI agents, not just data pipelines. The project crossed 230,000 active users in late 2025 — a 141% increase in one year — backed by $180M in funding led by Accel at a $2.5 billion valuation. Over 34% of Fortune 500 companies now use n8n enterprise features, and the platform serves 3,000+ enterprise customers. If you’ve outgrown Zapier’s task-based pricing or want to own your automation infrastructure, n8n is the right starting point.

Why Developers Choose n8n Over Alternatives

n8n sits at the intersection of no-code simplicity and developer extensibility. You get 400+ native integrations covering every major SaaS tool, plus an HTTP node that connects to any REST API without a dedicated connector. The Community Edition is free forever with unlimited executions — a stark contrast to Zapier’s per-task billing that becomes expensive at scale. Self-hosting on a VPS costs $3–7/month, giving you full data control and no vendor lock-in.

n8n vs. Zapier vs. Make: Which Should You Start With in 2026?

n8n is the right choice for developers who want unlimited customization, data ownership, and AI agent capabilities — Zapier wins on simplicity for non-technical users, while Make sits in the middle. Here is what the comparison looks like in practice for a technical beginner in 2026. Zapier has 7,000+ integrations compared to n8n’s 400+ native ones, but n8n’s HTTP node effectively closes that gap for any REST-accessible service. Zapier’s pricing starts free but scales steeply — at 2,000 tasks/month you’re already paying $49/month. n8n Cloud’s Starter plan is €24/month for 2,500 executions, and self-hosting is free. Make (formerly Integromat) offers a visual scenario builder similar to n8n but lacks the native AI agent nodes that n8n 2.0 introduced. The clear verdict for developers: start with n8n if you care about AI workflows, self-hosting, or cost at scale.

When to Choose Zapier Instead

Zapier remains the better choice if you are a non-technical user who needs a working automation in five minutes and does not plan to write any code. Its natural language workflow builder and polished onboarding outperform n8n for simple use cases like “when I get a Gmail, create a Trello card.” For developers, Zapier’s limitations quickly become frustrating.

Setting Up n8n: Cloud Trial vs. Self-Hosting with Docker

n8n offers two paths for getting started: the n8n Cloud free trial (no setup required, browser-based in minutes) and self-hosting via Docker on a VPS or local machine. The cloud trial is the fastest way to try n8n — sign up at n8n.io and you get a 14-day free trial with full access to all features including AI nodes. For production use or privacy-sensitive workflows, self-hosting is the better long-term choice and costs as little as $3–7/month on a VPS like Hetzner or DigitalOcean. The self-hosted Community Edition has no execution limits, no feature gates, and stores all data in your own database. Most beginners should start with Cloud to learn the interface, then migrate to self-hosting once they understand their workflow volume and data sensitivity requirements. Docker setup takes under 10 minutes if you already have Docker installed, making n8n one of the easiest developer tools to self-host in 2026.

Option A: n8n Cloud (Recommended for Beginners)

1. Go to n8n.io and click Get started free
2. Create an account with your email
3. Your n8n instance is ready at your-instance.n8n.cloud
4. No Docker, no server configuration required

Option B: Self-Hosting with Docker

If you have Docker installed locally or on a VPS, run this single command:

docker run -it --rm \
  --name n8n \
  -p 5678:5678 \
  -v ~/.n8n:/home/node/.n8n \
  docker.n8n.io/n8nio/n8n

Then open http://localhost:5678 in your browser. For production, add a reverse proxy (Caddy or Nginx) with HTTPS and set these environment variables:

N8N_BASIC_AUTH_ACTIVE=true
N8N_BASIC_AUTH_USER=your_username
N8N_BASIC_AUTH_PASSWORD=your_password
WEBHOOK_URL=https://your-domain.com

For persistent storage and auto-restart, use Docker Compose with a PostgreSQL database instead of the default SQLite.

Docker Compose for Production

version: "3.8"
services:
  n8n:
    image: docker.n8n.io/n8nio/n8n
    restart: always
    ports:
      - "5678:5678"
    environment:
      - DB_TYPE=postgresdb
      - DB_POSTGRESDB_HOST=postgres
      - WEBHOOK_URL=https://your-domain.com
    volumes:
      - n8n_data:/home/node/.n8n
  postgres:
    image: postgres:15
    environment:
      POSTGRES_DB: n8n
      POSTGRES_USER: n8n
      POSTGRES_PASSWORD: your_password
    volumes:
      - postgres_data:/var/lib/postgresql/data

n8n Core Concepts: Nodes, Triggers, Actions, and Workflows Explained

n8n workflows are built from nodes — individual building blocks that each perform one function, connected by edges that pass data between them. Every workflow starts with a trigger node that defines when the workflow runs: on a schedule, when a webhook fires, when an email arrives, or manually. Action nodes follow: they read from Google Sheets, send a Slack message, call an OpenAI API, or write to a database. Data flows between nodes as JSON objects, and you can transform that data using n8n’s expression language ({{ $json.fieldName }}) or a Code node for full JavaScript/Python. The canvas editor shows your entire workflow visually, and the execution log shows exactly what data passed through each node — making debugging straightforward even for beginners.

Key Node Types You’ll Use Daily

• Trigger nodes: Schedule (cron), Webhook, Gmail Trigger, Slack Trigger, Manual Trigger
• Core action nodes: HTTP Request, Code (JS/Python), Set (transform data), If (branch logic), Split In Batches
• App nodes: Google Sheets, Gmail, Slack, Notion, Airtable, Postgres, MySQL
• AI nodes: AI Agent, OpenAI, Anthropic, Google Gemini, LangChain chain, Memory nodes
• Utility nodes: Merge, Filter, Sort, Limit, Wait, Error Trigger

How Data Flows Between Nodes

Each node outputs an array of items. The next node receives that array and processes each item. If a node outputs 10 items, the next node runs 10 times — once per item. You reference data with expressions like {{ $json.email }} (current item’s email field) or {{ $('NodeName').item.json.fieldName }} (data from a specific upstream node).

Build Your First n8n Workflow (Step-by-Step: Schedule + Google Sheets)

Your first n8n workflow will run on a schedule, fetch data from a public API, and write it to Google Sheets — a pattern that covers 80% of real automation needs and teaches you the three core node types every n8n workflow uses: Trigger, Transform, and Action. This complete walkthrough takes about 15 minutes and requires only an n8n account and a Google account. By the end, you will have a live automation that runs every morning without any manual intervention. The workflow tracks Bitcoin prices daily, but the exact same pattern applies to monitoring stock prices, competitor website changes, weather data, or any REST API that returns JSON. Learning this pattern first gives you the foundation to understand every more complex workflow you’ll build afterward — including the AI agent workflow in the next section.

Step 1: Create a New Workflow

1. Open your n8n instance and click New Workflow (top right)
2. Name it “Daily Price Tracker” (double-click the title to rename)
3. Click the + button on the canvas to add your first node

Step 2: Add a Schedule Trigger

1. Search for “Schedule” and select Schedule Trigger
2. Set Trigger Interval to Days and Days Between Triggers to 1
3. Set the time to 08:00 in your timezone
4. Click the X to close the panel — the node is configured

Step 3: Fetch Data with HTTP Request

1. Click + after the Schedule Trigger node
2. Search for “HTTP Request” and select it
3. Set URL to a public API endpoint (example: https://api.coindesk.com/v1/bpi/currentprice.json for Bitcoin price)
4. Set Method to GET
5. Click Test Step — you should see JSON data in the output panel

Step 4: Transform the Data with a Set Node

1. Add a Set node after HTTP Request
2. Add a field: Name = price, Value = {{ $json.bpi.USD.rate }}
3. Add a field: Name = date, Value = {{ $now.toFormat('yyyy-MM-dd') }}
4. Add a field: Name = currency, Value = USD

Step 5: Write to Google Sheets

1. Add a Google Sheets node
2. Click Create new credential → authenticate with your Google account
3. Set Operation to Append or Update Row
4. Select your spreadsheet and sheet
5. Map your columns: Column A = {{ $json.date }}, Column B = {{ $json.price }}
6. Click Test Step — a new row should appear in your spreadsheet

Step 6: Activate the Workflow

Toggle the Active switch at the top of the canvas. Your workflow now runs every morning at 8 AM automatically.

Build Your First AI Workflow in n8n (Using the AI Agent Node)

n8n’s AI Agent node — introduced in n8n 2.0 with native LangChain integration — lets you build autonomous agents that reason, use tools, and handle multi-step tasks without hardcoding every logic branch. This tutorial builds an AI email responder: it reads incoming emails, drafts context-aware replies using an LLM, and saves the drafts back to Gmail. You can build a working AI chatbot connected to Telegram, Slack, or Discord in under an hour using the same approach. The AI Agent node is n8n’s most powerful feature for 2026 — it makes workflows dynamic in a way that traditional automation cannot match. Unlike a simple OpenAI API call in a Code node, the AI Agent node gives the LLM access to tools (like HTTP requests, Google Sheets reads, or Slack sends) and lets it decide which tools to use and in what order based on the input. This is the architecture behind production AI assistants. You wire the agent’s goal into the System Prompt, attach the tools you want it to use, and n8n handles the reasoning loop — including retry logic when the agent’s first tool call doesn’t produce the expected result.

Step 1: Create a New Workflow with Gmail Trigger

1. Create a new workflow named “AI Email Responder”
2. Add a Gmail Trigger node
3. Set Event to Message Received
4. Filter by label or sender if needed (e.g., only emails with subject containing “Support”)
5. Connect your Gmail credential

Step 2: Add the AI Agent Node

1. Add an AI Agent node after Gmail Trigger
2. Set Prompt (the agent’s system context):
3. Under Chat Model, click + and add an OpenAI Chat Model node
4. Select gpt-4o-mini (cost-effective for email drafting)
5. Enter your OpenAI API key

Step 3: Add Memory (Optional but Recommended)

1. Connect a Window Buffer Memory node to the AI Agent’s Memory input
2. Set Session ID to {{ $json.threadId }} so the agent remembers conversation context per thread
3. Set Context Window Length to 10 messages

Step 4: Save the Draft to Gmail

1. Add a Gmail node after AI Agent
2. Set Operation to Create Draft
3. Set To to {{ $('Gmail Trigger').item.json.from.value[0].address }}
4. Set Subject to Re: {{ $('Gmail Trigger').item.json.subject }}
5. Set Email Type to Text
6. Set Message to {{ $json.output }} (the AI Agent’s response)

Step 5: Test and Activate

Send yourself a test email, then click Test Workflow in n8n. You should see a draft appear in your Gmail Drafts folder within seconds. Once verified, activate the workflow.

5 Beginner-Friendly n8n Workflow Templates to Clone and Customize

n8n has 9,600+ community workflow templates available for free at n8n.io/workflows — you can import any template into your instance with one click and start customizing immediately. The best beginner templates are complete, working examples with real credentials you can swap in. These five templates cover the most common beginner use cases: RSS monitoring, Slack notifications, spreadsheet automation, AI summarization, and webhook-based data capture. Each template teaches a different pattern that you’ll reuse across dozens of future workflows. Cloning templates is faster than building from scratch and teaches you how experienced n8n builders structure their nodes — including error handling and branching logic you might not think to add on your own. To import a template, navigate to the Templates tab in n8n, search by keyword, and click “Use for free” to load it directly into your workflow editor with all node configurations pre-filled.

Template 1: RSS Feed to Slack Notifier

• What it does: Monitors an RSS feed every hour and posts new items to a Slack channel
• Nodes used: Schedule Trigger → RSS Read → If (check if new) → Slack
• Why it’s great for beginners: Teaches the Schedule + If + Action pattern with no external APIs needed
• Clone it: Search “RSS Slack” in the n8n template library

Template 2: Google Sheets Row to Email

• What it does: When a new row is added to a Google Sheet, sends a formatted email via Gmail
• Nodes used: Google Sheets Trigger → Set → Gmail
• Why it’s great: Shows how to trigger from data changes, not just schedules

Template 3: AI Blog Post Summarizer

• What it does: Takes a URL from a form, scrapes the content, summarizes it with GPT-4o, saves to Notion
• Nodes used: Webhook → HTTP Request → OpenAI → Notion
• Why it’s great: Demonstrates the full LLM integration pattern

Template 4: Telegram AI Chatbot

• What it does: A Telegram bot that answers questions using an AI Agent with web search tools
• Nodes used: Telegram Trigger → AI Agent → Telegram (send reply)
• Why it’s great: Shows real-time chat integration with a stateful AI agent

Template 5: GitHub Issue to Linear Task

• What it does: When a GitHub issue is created with a specific label, creates a corresponding Linear task
• Nodes used: GitHub Trigger → If → Linear
• Why it’s great: Pure app-to-app integration — no AI required, shows webhook routing

Common n8n Mistakes Beginners Make (And How to Avoid Them)

Most beginner n8n errors fall into five categories: credential misconfiguration, misunderstanding data flow, overcomplicating the first workflow, ignoring error handling, and not testing incrementally. Understanding these pitfalls before you hit them saves hours of debugging. The most common mistake is building a 15-node workflow before testing any individual node — n8n lets you test each node independently, which is the right approach. Another frequent error is referencing data incorrectly: $json.field works inside the current node’s expression editor, but you need $('NodeName').item.json.field to reference a specific upstream node’s output. Beginners also frequently forget that n8n processes arrays of items — if your HTTP node returns a list of 50 records, the next node runs 50 times, which can accidentally trigger 50 Slack messages instead of one summary message. Learning to use the Merge node and Split In Batches node early prevents these data flow surprises in production workflows.

Mistake 1: Not Testing Each Node Individually

Wrong approach: Build the entire workflow, then run it end-to-end and debug the error.
Right approach: After adding each node, click Test Step to verify its output before adding the next one.

Mistake 2: Incorrect Data References

Wrong: {{ $json.email }} in a node three steps downstream (may reference the wrong node)
Right: {{ $('Gmail Trigger').item.json.from.value[0].address }} — always be explicit about the source node

Mistake 3: Missing Error Handling

Add an Error Trigger workflow for production automations. When a node fails, n8n can route to an error handler that sends you a Slack message or logs the failure to a spreadsheet instead of silently dropping data.

Mistake 4: Hardcoding Credentials in Expressions

Never put API keys directly in node expressions. Always use n8n’s Credentials system — it encrypts secrets and makes rotating them trivial.

Mistake 5: Ignoring Rate Limits

When processing large lists, use the Split In Batches node to process items in groups of 10–50 instead of hammering APIs with 1,000 parallel requests.

n8n Pricing Breakdown: Free Community Edition vs. Cloud Plans

n8n’s pricing model is developer-friendly compared to competitors: the Community Edition is permanently free with no execution limits, while cloud plans serve teams that want managed infrastructure. For most developers starting out, the Community Edition on a $5/month VPS is the best value — you get full features, unlimited runs, and complete data ownership. The workflow automation market is projected to reach $37.45 billion by 2030 with a 9.52% CAGR, meaning the investment in learning n8n pays off as automation skills become standard for developers.

Is the Community Edition Really Free?

Yes — n8n’s Community Edition is MIT-licensed for self-hosted use. There are no execution caps, no feature tiers, and no credit card required. The only limitation is that you manage your own infrastructure: updates, backups, and uptime. For a personal automation server, this is a non-issue.

What to Build Next: Intermediate n8n Workflows for Growing Developers

Once you’ve built your first two or three workflows, the natural progression is toward AI agents that can use multiple tools, workflows that span multiple services in complex ways, and sub-workflows that you reuse across projects. n8n’s 2026 MCP (Model Context Protocol) integration lets you connect AI agents directly to external tools and services using the emerging standard, making n8n one of the first automation platforms to support agent-to-agent communication via MCP. The key intermediate skill is sub-workflow architecture: instead of one 30-node monolith, build modular workflows that call each other — each does one thing and does it well. This architecture is also what allows you to run the same sub-workflow from multiple parent workflows, maintaining DRY principles in your automation stack. At the intermediate level you will also want to learn n8n’s built-in error handling, execution monitoring, and queue mode for high-volume production workflows.

Intermediate Project Ideas

1. Multi-source lead enrichment: When a new lead lands in HubSpot, trigger a sub-workflow that queries LinkedIn, Clearbit, and your own database, then scores the lead and routes it to the right sales rep
2. AI customer support router: Gmail Trigger → AI Agent classifies intent → routes to different Slack channels or ticket queues based on category
3. Automated social media scheduler: Pull approved content from a Notion database → format for each platform → schedule posts via Twitter/LinkedIn/Bluesky APIs
4. Code review assistant: GitHub PR webhook → AI Agent reads the diff → posts a structured review comment with specific suggestions
5. Real-time inventory monitor: Scheduled webhook polls your e-commerce API every 15 minutes → alerts Slack when stock drops below threshold → triggers a reorder email

Key Skills to Learn Next

• Sub-workflows: Use Execute Workflow node to call reusable workflow modules
• Custom nodes: Build TypeScript nodes for internal APIs using n8n’s node creator CLI
• n8n + MCP: Connect AI agents to external context sources using the Model Context Protocol
• Queue mode: Enable Redis-backed queue processing for high-throughput workflows
• Monitoring: Set up n8n’s built-in execution log with external alerting via Prometheus or Grafana

FAQ

The most common questions from developers just starting with n8n center on pricing, setup requirements, and how it compares to Zapier and Make. These answers reflect n8n’s 2026 feature set — including the n8n 2.0 AI Agent node, MCP integration, and the free Community Edition that makes n8n the most accessible automation platform for self-hosting developers. If you are evaluating whether n8n fits your use case, these five questions cover the decisions every beginner faces before writing their first workflow. The short summary: n8n is free to self-host with unlimited executions, works without coding experience for most tasks, connects to OpenAI and Anthropic natively through dedicated credential-secured nodes, beats Zapier on price and customization flexibility for technical users, and takes 20–45 minutes to go from zero to a working AI workflow on the n8n Cloud free trial.

Is n8n free to use?

Yes. n8n’s Community Edition is permanently free for self-hosted use with no execution limits. You pay only for your server costs (typically $3–7/month on a VPS). n8n Cloud starts at €24/month for teams who prefer managed infrastructure.

Can I use n8n without coding experience?

You can build most workflows without writing any code — the visual editor handles connections, and nodes have configuration forms. However, n8n gives you optional JavaScript and Python access in Code nodes for advanced transformations. If you know basic coding, you’ll get significantly more out of n8n than a zero-code user.

How does n8n connect to AI models like ChatGPT?

n8n 2.0 includes native OpenAI, Anthropic, and Google Gemini nodes. You add your API key once in n8n’s credentials vault, then drag an OpenAI node into any workflow and reference the credential. The AI Agent node uses these models to power autonomous reasoning with tool use.

What’s the difference between n8n and Zapier?

The key differences are: (1) n8n is open-source and self-hostable, Zapier is SaaS-only; (2) n8n’s Community Edition is free with unlimited runs, Zapier’s free tier allows only 100 tasks/month; (3) n8n has native AI agent nodes, Zapier does not; (4) n8n lets you write JavaScript/Python in-workflow, Zapier does not. Zapier wins on ease of use for non-technical users and has 7,000+ integrations vs. n8n’s 400+.

How long does it take to build a working AI workflow in n8n?

A simple AI workflow — like a Telegram chatbot or email responder — takes 20–45 minutes for a developer who is new to n8n. Complex multi-agent workflows with custom tools take a few hours. The fastest path is to start from a community template and customize it rather than building from scratch.