How to Use n8n: The Beginner's Guide (Cloud vs Self-Host, First Workflow, AI Node, and the Paid Line)

CodexFuse Team · 2026-10-29T00:56:17.946Z · 14 min read

If you want to learn how to use n8n, start with one small workflow, not with the whole platform. n8n is a visual automation tool: you connect apps and steps on a canvas, and n8n runs them for you on a schedule or when something happens, such as a form being submitted. This n8n for beginners guide walks you through the complete first-use system in order: choosing Cloud or self-hosting, building your first workflow, adding an AI node, keeping your credentials safe, handling errors, and understanding the paid line.

The core idea is simple, and it is the same idea every later workflow reuses. A workflow is a chain of nodes, and each node does one job. A trigger node decides when the workflow starts, the nodes after it fetch, reshape, or send data, and the last node delivers the result.

If you plan to self-host, read our best VPS for n8n guide before you pick a server, because the machine you run n8n on is part of the decision, not an afterthought.

How do I use n8n for the first time?

To use n8n for the first time, sign up for an n8n Cloud trial or install the free self-hosted Community edition, then create a workflow, add a trigger, add one node at a time, test each step as you go, and publish the workflow only when it works end to end. That loop is the whole beginner system, and every section below is one step of it.

1. Choose Cloud or self-hosting: pick Cloud if you want to start straight away, and self-hosting if you are comfortable running a server.
2. Create a workflow and give it a clear name, such as New enquiry alert, so you can find it again in a month.
3. Add a trigger first, and answer one question before anything else: how should this workflow start, on a schedule, or when something happens?
4. Add nodes one at a time, testing each node with Execute step before you add the next one.
5. Run the whole workflow with Execute workflow, check the real result, and only then select Publish so the trigger runs it for you.
6. Watch the Executions list every day for the first week, so a silent failure never sits unnoticed for days.

What is n8n, and what does a workflow actually do?

n8n is a workflow automation tool that connects apps and services without you writing a full program for each connection. You build on a canvas by dragging nodes into place and drawing lines between them, and data flows from one node to the next. A workflow is the saved chain itself, while an execution is one actual run of that chain, and keeping those two words apart clears up most early confusion.

Three node types cover almost every beginner workflow. A trigger starts the run, such as a schedule trigger that fires every morning or a form trigger that fires when someone submits a form. A processing node reshapes data, such as Edit Fields, which renames fields and turns raw output into a clean one-line report.

An action node delivers the result, such as sending an email, posting a message, or inserting a row into a data table. Once you can name the trigger, the processing, and the action in a workflow, you can read any n8n workflow, including ones other people built.

The other idea beginners need early is expressions, which is how data moves between nodes. The output of one node becomes the input of the next, and you can drag a field from a previous node into a value box instead of typing a fixed value. That is how a message template like "New enquiry from" can carry a different name on every run, without you editing the workflow each time.

Should I use n8n Cloud or self-host n8n?

Choose n8n Cloud if you want to start in minutes with no server to look after, and choose self-hosting if you want full control and are comfortable managing a server. This is the one decision to get right first, because it sets your cost, your maintenance, and your limits. n8n itself recommends Cloud for people who want to start right away or who do not have technical expertise, and self-hosting for expert users who can secure and maintain a server.

Self-hosted n8n has no execution cap imposed by n8n, which is the mirror image of Cloud pricing, where executions are the thing you buy. That does not make self-hosting free in practice: the server still costs money, and your time maintaining it is real. The honest comparison is Cloud's monthly price against a server's monthly price plus the hours you spend keeping it updated and backed up.

If nobody on your team manages servers, Cloud wins that comparison, and the self-host route can wait until you have a reason for it.

One more distinction matters before you choose. The self-hosted Community edition is free with almost the complete feature set, and registering it adds folders, debug in the editor, and custom execution data. Paid Cloud plans add managed hosting and team features on top, not a different automation engine.

So the question is never which version automates better; it is who looks after the machine, and what happens when something breaks at an inconvenient hour.

How do I set up self-hosted n8n with Docker?

To set up self-hosted n8n, install Docker, start n8n in a container with a persistent volume, and open it in your browser at port 5678. n8n's own docs recommend Docker for most self-hosting needs, because it keeps n8n isolated from the rest of your system and makes updates simpler. For a real server, the docs point to Docker Compose as the recommended route; the single command below is the quick local start that proves the setup works.

First, learn what VPS hosting is if the word server still feels vague, because a VPS is the usual home for a self-hosted n8n install. Then, if you have never run a container before, our Docker beginner guide covers installing Docker and the five commands that matter. With Docker installed, the docs' own start commands are: create a volume to hold your data, then run the n8n image with your timezone set and port 5678 exposed.

• docker volume create n8n_data creates the persistent volume, so your workflows and credentials survive a container restart.
• docker run -it --rm --name n8n -p 5678:5678 -e GENERIC_TIMEZONE="Your/Timezone" -e TZ="Your/Timezone" -v n8n_data:/home/node/.n8n n8nio/n8n starts n8n, replacing Your/Timezone with your real timezone, such as America/New_York.
• Open http://localhost:5678 in your browser to reach the n8n editor and create your account on first run.

Two warnings belong in every self-hosting guide, and n8n's docs state both plainly. Self-hosting requires technical knowledge, including securing the server, and mistakes can lead to data loss, security issues, and downtime. Never expose an n8n editor to the open internet without authentication and HTTPS, and back up the data volume before you update, because an update you cannot roll back is a risk you chose, not bad luck.

How do I build my first workflow in n8n?

Build your first workflow as a small chain with one trigger, one data fetch, one reshape, and one saved result, and test it after every node. The example below follows the tutorial pattern n8n's own guides teach: a morning report that runs on a schedule, fetches data, cleans it up, and saves it. It is deliberately boring, because a boring workflow that works teaches you more than an ambitious one that fails in three places at once.

1. Add a Schedule Trigger node and set it to run every day at 8:00 AM, so the workflow has a clear starting point.
2. Add an HTTP Request node after the trigger and point it at the data source you want, such as a weather endpoint or another web service with an address you control.
3. Add an Edit Fields node and keep only the fields your report needs, such as temperature, condition, and location, renaming anything cryptic into plain words.
4. Add a Data table node with an Insert row action, and map each cleaned field into its own column, so every run leaves a record you can check.
5. Select Execute step after each node to see that node's real output before moving on, instead of guessing what it produced.
6. Select Execute workflow to run the whole chain once, confirm the new row appeared, and only then select Publish so the schedule runs it for you.

The habit inside those steps matters more than the example itself. Test after every node, because a workflow with five untested nodes has five possible failure points and no clue which one broke. Name the workflow clearly, and keep each workflow small: one job per workflow, several small workflows instead of one giant one.

When a workflow grows past a handful of nodes, split it, because a chain you cannot explain in one sentence is a chain you cannot debug at midnight.

One trap catches nearly every beginner here. When you build a second version of a workflow, unpublish the first one, or both will run.

n8n gives exactly that warning in its own guide, and duplicate runs are the classic cause of double emails, double rows, and double charges from downstream services. Publish is a switch, not a save button: saving keeps your draft, publishing turns the trigger on.

How do I add an AI node to an n8n workflow?

To add an AI node, add an AI Agent node to your workflow, connect it to a model through credentials you add in n8n, give it the tools it may use, and test it with real inputs before it ever messages a customer. The AI node is not a separate product: it is a node like any other, sitting in the middle of a chain, reading the previous node's output and handing its answer to the next node. That placement is the whole point, because the AI step inherits the same testing discipline as every other step.

The most useful first AI workflow is support message triage, the pattern business guides keep returning to. A new email or chat message arrives, the AI classifies it by type, such as sales, complaint, or delivery, and detects urgency, then the workflow routes it to the right person and logs it in a sheet for weekly review. The AI does the reading and sorting; a human still handles the reply.

That boundary is what keeps an AI workflow trustworthy, because the model's mistakes land in a queue a person checks, not in a customer's inbox.

Budget for the AI step separately from n8n itself. On n8n Cloud's Starter plan, agents, MCP, and AI nodes are included, along with AI models without needing your own API keys. On self-hosted n8n, you add your own model provider credentials in n8n's credential store, and the model usage is billed by that provider.

Never paste a real API key into a node field or a message template: credentials belong in n8n's credential store, where they are saved once and reused safely, and a key sitting in a workflow's text is a leak waiting for a screenshot.

Never let an AI node send messages to customers without testing its edge cases first. Feed it angry messages, vague messages, messages in another language, and messages with missing details, and watch what it produces for each. If any answer would embarrass you in front of a customer, add a human review step before the sending node, and keep it there until the failure rate earns its removal.

How much does n8n cost, and when do I pay?

n8n costs nothing for the self-hosted Community edition apart from your server, and its Cloud plans start at 20 euros per month billed annually, with a 14-day trial that needs no credit card. Prices below were checked on n8n's own pricing page on October 28, 2026, and Cloud prices can change, so re-check the pricing page before you commit. The Cloud model is worth understanding before the numbers, because it differs from most tools: n8n charges for completed workflow runs, not for steps or users, and its own pricing page puts it bluntly, that if a run does not finish, you do not pay for it.

The paid line for a beginner is therefore simple. Start on the Cloud trial or the free Community edition, and pay only when your workflows are running reliably and you have outgrown the free route's limits. If your automations are a handful of daily reports and alerts, Starter's execution allowance covers them with room to spare.

If you are automating a whole business, with many workflows firing all day, Pro or Business is a production decision, and it should follow measured execution counts from your Executions list, not a guess made on day one.

For the self-host route, compare honestly against Cloud before you commit. Our cheap VPS hosting guide works through the real total cost of a small server, including the renewal price, which is the number that matters after the first term. A small VPS often costs less per month than Cloud Starter, but the gap closes fast once you count the hours you spend updating, backing up, and fixing it.

The cheapest option is the one you will actually maintain.

n8n vs Zapier: which one should a beginner pick?

Pick Zapier if you want the simplest possible automation with no server and no learning curve, and pick n8n if you want self-hosting, deeper control, and room to grow into code. Zapier is the no-code lane: pick a trigger app, pick an action app, and you are done, which is why it suits a first automation built in an afternoon. n8n asks a little more of you up front, and repays it with a visual canvas you can inspect node by node, a self-hosted edition with no execution cap, and escape hatches like HTTP requests and expressions when an app's ready-made node does not do quite what you need.

Our Zapier beginner guide covers that lane in full, and it is worth reading even if you choose n8n, because the trigger-and-action mental model is identical. The real difference is not the first workflow, which both tools handle in minutes, but the tenth workflow. In Zapier, the tenth workflow costs more steps and more money on a plan ladder.

In n8n, the tenth workflow costs more of your attention: more nodes to test, more credentials to manage, and more executions to watch. Choose the cost you would rather pay.

There is also a middle path beginners miss. You can start on n8n Cloud, learn the canvas and the testing habits, and move to self-hosting later, when a server makes sense and you have workflows worth protecting. Nothing about starting on Cloud locks you out of self-hosting, and nothing about self-hosting requires you to start there.

The skill that transfers is not the hosting choice; it is building small, tested, named workflows with error handling from the start.

What goes wrong for beginners, and how do I fix it?

Most beginner failures in n8n come from four mistakes, and all four are avoidable once you know their names. The first is automating a process that is still unclear or changing every week: if you cannot describe the manual process on paper, step by step, the workflow will encode your confusion, not remove it. The second is building one huge workflow instead of several small ones, which turns every failure into a mystery.

The third is storing passwords in plain text instead of using n8n credentials, which turns a shared screenshot into a security incident. The fourth is skipping error alerts, so failures sit unnoticed for days.

Fix errors with a system, not with hope. Add an Error Trigger node, which starts a separate workflow that alerts you when another workflow fails, so a broken chain tells you itself instead of waiting for a customer to complain. Check the Executions list daily for the first week of any new workflow, and read the failed runs, not just the successful ones.

When a node fails, run it alone with Execute step and look at its actual input and output, because the input is usually the surprise, not the node's settings.

Finally, keep a short checklist before you publish anything that touches real people or real money. Test with real data, not just the sample that happened to work. Confirm the workflow sends to the right channel, the right sheet, and the right people.

Confirm the second version is the only published version. And confirm you know how to turn it off: unpublishing a workflow is the fastest fix for any bad run, and it should take seconds, not a support ticket.

Sources

• Lilach Bullock, "How to Use n8n: Build Your First Workflow (Beginners)" (published October 23, 2026) — Cloud trial terms, the Starter plan's 2,500 executions, the Registered Community extras, the 15-minute quickstart claim, and the publish-and-unpublish warning. https://www.lilachbullock.com/how-to-use-n8n/
• n8n, "Plans and Pricing" (checked October 28, 2026) — Starter 20 euros per month, Pro 50 euros per month, and Business 667 euros per month, all billed annually, no credit card required for the trial, and the rule that you pay only when a workflow runs from start to finish. https://n8n.io/pricing/
• n8n Docs, "Install with Docker" (checked October 28, 2026) — Docker as the recommended route for most self-hosting, the persistent volume and port 5678, the Docker Compose recommendation for servers, and the warning that self-hosting needs server expertise. https://docs.n8n.io/hosting/installation/docker/
• SofThat, "n8n Workflow Automation: A Beginner's Guide for Businesses" (checked October 28, 2026) — the seven-step first-workflow method, the four common beginner mistakes, and the Error Trigger approach to failure alerts. https://softhat.pk/blog/n8n-workflow-automation-guide/

Frequently asked questions

Is n8n free?

The self-hosted Community edition is free, with almost the complete feature set, but you still pay for the server it runs on and the time you spend maintaining it. n8n Cloud is a paid, managed service with a 14-day trial that needs no credit card, and its Starter plan costs 20 euros per month billed annually.

Do I need to know how to code to use n8n?

No. n8n's own quickstart says a first workflow takes about 15 minutes and needs no API keys or coding experience, because you connect nodes on a canvas and drag fields between them instead of writing code. Code helps later, when you want custom logic, but it is not the price of admission.

What is the difference between n8n Cloud and self-hosted n8n?

n8n Cloud is hosted and managed by n8n, so you pay a monthly price based on executions and maintain nothing. Self-hosted n8n runs on your own server, costs nothing to n8n, has no execution cap from n8n, and leaves install, updates, backups, and security entirely to you.

How do I make an n8n workflow run automatically?

Test it with Execute workflow until it works end to end, then select Publish in the canvas header, and the trigger runs it for you from then on. If you later build a second version, unpublish the first one first, or both versions will run.

Should a beginner choose n8n or Zapier?

Choose Zapier if you want the simplest no-code automation with no server to manage, and choose n8n if you want self-hosting, deeper control over each step, and room to grow into code and AI nodes. The trigger-and-action idea is the same in both, so learning one teaches you the other.