What is n8n used for? The recurring jobs teams hand to it, each with a worked example of trigger, steps and systems, plus the self-hosting reality.
What is n8n used for? n8n is a fair-code, node-based workflow tool that connects apps and APIs on a visual canvas. Teams use it to sync data between systems, send alerts, run CRM and ecommerce flows, triage email and support, build reports, scrape web pages and wire up AI workflows. Its own site aims it at technical teams in IT, security, dev and sales.
What is n8n and what do people build with it?
The docs describe n8n as “a fair-code licensed workflow automation tool that combines AI features with business process automation” (docs.n8n.io). The source code is public and you can run it yourself, but n8n’s license page states that “according to the Open Source Initiative (OSI), open source licenses can’t include limitations on use, so we do not call ourselves open source” (fair-code license).
The n8n homepage pitches “AI agents and workflows for technical teams” and names a job per function: IT Ops can “On-board new employees”, Sec Ops can “Enrich security incident tickets”, Dev Ops can “Convert natural language into API calls”, Sales can “Generate customer insights from reviews”.
Strip the marketing and the recurring work falls into seven buckets, each with a worked example below.
How does n8n actually work?
Everything in n8n is a workflow, and a workflow is a graph of nodes. The glossary puts it plainly: “An n8n workflow is a collection of nodes that automate a process. Workflows begin execution when a trigger condition occurs, when you run them manually, or when another workflow calls them. They execute node by node, and can branch, merge, and loop” (n8n glossary). The n8n.io integrations directory lists 2,366 integrations.
Triggers, actions and the canvas
A node does one of two things. Per the node types reference, triggers “start a workflow in response to specific events or conditions in your services”, while actions “manipulate data, perform operations on external systems, and trigger events in other systems”. The Schedule Trigger runs workflows “at fixed intervals and times”, and the Webhook node lets you “trigger workflows from services that don’t have a dedicated app trigger node”. You wire them left to right on the canvas, and If, Switch, Merge and Loop Over Items handle branching and repetition (flow logic).
Code node and HTTP Request node
The Code node “lets you run your own JavaScript or Python inside a workflow, so you can transform data or add logic the built-in nodes don’t cover”. The HTTP Request node “allows you to make HTTP requests to query data from any app or service with a REST API”, built by hand or by importing a curl command.
What is n8n used for? The jobs teams hand to it
Grouped by business function, each example names the trigger, the steps and the systems touched.
See the jobs teams hand to an AI employee →
Data sync between apps
The job: keep the CRM, billing and project tools telling the same story. Example: a HubSpot Trigger fires when a deal changes stage, an If node checks for “Closed Won”, an Edit Fields node reshapes the deal into the finance sheet’s columns, a Google Sheets node appends the row and a Notion node creates the client page. Systems touched: HubSpot, Google Sheets, Notion.
Notifications and alerts
The job: push an event to where someone will see it. Example: a Stripe Trigger receives a Stripe webhook event, for example a failed invoice payment, an If node keeps invoices above a threshold, and a Slack node posts customer, amount and link into #billing while a Gmail node emails the account owner. Systems touched: Stripe, Slack, Gmail.
CRM and ecommerce workflows
The job: enrich leads and process orders without anyone copying fields. Example: a Shopify Trigger receives the order webhook, an If node separates first-time from repeat buyers, a HubSpot node creates or updates the contact and deal, and a Slack node drops a one-line summary in #orders. Systems touched: Shopify, HubSpot, Slack.
Email and support triage
The job: route, tag and deduplicate the inbox. Example: a Gmail Trigger polls for new messages, a Switch node routes by sender domain and subject keywords, a Zendesk node creates the ticket, and a Remove Duplicates node drops items “identical to or surpassed by items seen in previous executions” so a customer chasing twice does not open two tickets. Systems touched: Gmail, Zendesk, Slack for the escalation ping.
Reporting
The job: pull numbers from several places on a schedule and land them in one view. Example: a Schedule Trigger fires Monday at 08:00, a Postgres node queries last week’s signups and revenue, a Google Sheets node reads the ad spend tab, a Summarize node aggregates them “in a manner similar to Excel pivot tables”, an HTML node renders the template and Gmail sends it. Systems touched: Postgres, Google Sheets, Gmail.
Web scraping
The job: watch pages that have no API. Example: a Schedule Trigger runs nightly, an HTTP Request node fetches a competitor’s pricing page, the HTML node’s “Extract HTML content” operation pulls plan names and prices, an Edit Fields node normalises them, a Google Sheets node appends a dated row, and an If node posts to Slack only when a value changed. Systems touched: the target site, Google Sheets, Slack. Check the target’s terms of use before scheduling this.
AI workflows
The job: put a model in the loop where rules are not enough. The docs say n8n lets you “connect different LLM providers such as OpenAI, Anthropic, and Google, add tools and memory, and combine several models in one workflow” (integrate AI), and the AI Agent node works like this: “Connect a chat model and one or more tools, and the agent decides which tools to call to complete a task.” Example: a Webhook receives a support message, the AI Agent node reads it with HubSpot and Zendesk nodes attached as tools, and the Slack node serves as “a human review step for AI Agent tool calls” before any reply goes out. Systems touched: a model provider, HubSpot, Zendesk, Slack.
Browse the 3,200+ integrations →
Self-hosting vs n8n Cloud: the honest “used for” answer
The hosting docs say “You can self-host n8n on your own infrastructure, on-premises, or in a private cloud, using Docker Compose, one-line setup, or other deployment methods” and “Without a license key, n8n runs as the free Community edition” (host n8n). The decision page spells out the trade: Cloud is “Fully managed, no setup or maintenance required”, while self-hosting “Requires setup (npm, Docker, or server)” and expertise is “Required for installation and configuration” (choose how to use n8n). Self-hosting means your team owns updates, backups, the database and the server bill.
n8n Cloud bills by execution. The pricing page states “All plans include unlimited users & workflows and every integration. Pricing based on monthly workflow executions, regardless of complexity”, with Starter at 20€ a month billed annually for 2,500 executions, Pro at 50€ a month billed annually for 10,000, Business at 667€ a month billed annually for 40,000 and self-hosted only, and Enterprise on request (n8n pricing). An execution “is a single run of your entire workflow. It doesn’t matter how many steps are in the workflow or how much data it processes.” The Cloud trial needs no credit card, and hosted data “is stored within the EU, on servers located in Frankfurt, Germany”. Pricing checked October 2026.
What the fair-code license allows
The Community edition is free, and the edition comparison says it “includes almost the complete feature set of n8n” but leaves out environments, external secrets, log streaming, projects, SSO, sharing of workflows and credentials, and Git version control, which sit in the paid Business and Enterprise plans. The Sustainable Use License is “a fair-code software license created by n8n in 2022” with three limits: use or modify the software “only for your own internal business purposes or for non-commercial or personal use”, distribute it “only if you do so free of charge for non-commercial purposes”, and keep the licensing notices intact. Running n8n for your own company’s work is in bounds; shipping n8n to customers inside your product needs a commercial agreement.
Who should use n8n, and when it is the wrong tool
Who | Fit | Why |
|---|---|---|
Solo marketers | Partial | The canvas is approachable, but self-hosting needs Docker and server skills, and Cloud starts at 20€ a month billed annually. |
Ops teams | Good, with a builder | The seven jobs above are ops work, but someone has to own the graph when an API changes. |
Developers | Strong | The Code node, the HTTP Request node and self-hosting are built for people who read API docs. |
AI builders | Strong, if hands-on | The AI Agent node, tool sub-nodes and MCP support cover a lot, but every tool call is still wired by hand. |
Signs you have outgrown it: nobody wants to maintain the server; the workflows change weekly and each change is a rebuild; the work needs judgement at every step; or the people asking for it do not want to see a canvas.
Other fits exist: Zapier’s pricing page offers a hosted Free plan with 100 tasks a month and two-step Zaps, and Make’s pricing page a Free plan at up to 1,000 credits a month; both are compared below. The head-to-head with an AI employee is in Viktor vs n8n and the shortlist in n8n alternatives; this page names no winner.
n8n vs Zapier vs Make
| n8n | Zapier | Make |
|---|---|---|---|
Source | Fair-code, source on GitHub, not called open source | Hosted by Zapier, no self-hosted option listed | Hosted by Make, no self-hosted edition listed |
Billing unit | Workflow executions, “regardless of complexity” | Tasks: “Every step in a Zap and every external connector call uses tasks” | Credits: “Each module action in your scenario” is one credit |
Free option | Self-hosted Community edition; Cloud trial, no credit card (n8n pricing) | Free plan, 100 tasks a month, two-step Zaps (Zapier pricing) | Free plan, up to 1,000 credits a month, 15-minute minimum interval between runs (Make pricing) |
Entry paid plan | Starter, 20€ a month billed annually, 2,500 executions (n8n pricing) | Professional, listed from $19.99 a month, multi-step Zaps (Zapier pricing) | Core, priced per 10,000 credits a month (Make pricing) |
Self-hosting | Yes: one-line setup, Docker Compose, npm or a cloud provider (install options) | Not offered | Not offered |
The billing unit is the real difference: one n8n execution covers a whole run, while Zapier and Make meter each step or module. n8n vs Zapier covers pricing, hosting and error handling in detail, and what Zapier is used for is the companion to this page.
Building the graph yourself vs describing the Task in plain language
Every example above is built the same way: pick a trigger, add nodes, map fields, test, publish, and come back when an API changes. The other way to get the Monday report or the order-to-CRM flow is to describe it once and let an AI employee work out the steps. Viktor lives in Slack and Microsoft Teams, connects to 3,200+ tools where “Most connectors are one-click OAuth”, runs recurring work as Tasks on a schedule, and keeps the sensitive steps gated: “Money moves, code pushes, and customer emails wait for your explicit approval in Slack”, as the security page puts it. The flow is the same; the canvas is gone.
Start with up to $100 in free credits. No credit card required.
FAQ
Is n8n completely free?
The self-hosted Community edition is free: per the pricing page, “A standard, self-hosted version of n8n is available on GitHub”. n8n Cloud is paid after the trial, from 20€ a month billed annually.
Can I self-host n8n or do I need n8n Cloud?
Both are supported for production. The decision page recommends Cloud when you “Want to start right away” or “Don’t have technical expertise”, and self-hosting when you “Need full control over deployment”.
Can n8n build AI workflows?
Yes. The AI Agent node connects a chat model and tools, and the docs cover chains, memory, vector stores and MCP servers. The Slack node can act as a human approval step before a tool runs.
Do I need to code to use n8n?
Not for the standard app nodes, which are configured with forms and drag-and-drop mapping. You will meet JavaScript quickly, because expressions that pull values from earlier nodes are, in the docs’ words, “small pieces of JavaScript-like code”, and the Code node is the usual answer when no built-in node covers a transformation.
Is n8n open source?
n8n says no: its license page states that OSI open source licenses “can’t include limitations on use, so we do not call ourselves open source”. The code is public and free to run for internal business use.
What counts as an execution on n8n Cloud?
One complete run of a workflow, however many nodes it contains. The executions docs add that only production executions, “started automatically by triggers, schedules, or polling”, count towards the plan quota; manual test runs in the editor do not.