← All tutorials
n8n β€” workflow automation

n8n β€” automate the glue between your apps

n8n is a workflow-automation tool: you wire together triggers and actions on a visual canvas to connect apps, APIs and databases β€” automating the repetitive "when X happens, do Y and Z" work that would otherwise be manual or bespoke scripts. It's source-available and self-hostable, with an escape hatch to real code when the no-code nodes aren't enough.

a trigger fires→ nodes transform & route data→ actions happen
One line. n8n turns "I keep manually copying data between these five tools" into a workflow that runs itself β€” visually built, code-extensible, and running on your infrastructure if you want it.

Why  The problem it solves

Modern work is spread across dozens of SaaS apps that don't talk to each other. Bridging them means either manual busywork (copy a lead from a form into the CRM, ping Slack, add a row to a sheet…) or custom glue scripts you have to write, host, monitor and maintain forever. Both scale badly.

n8n collapses that into declarative workflows: connect a trigger to a chain of actions, and it handles the scheduling, the API auth, the retries and the data plumbing. It targets three pains at once:

πŸ” Repetitive tasks

Anything you do the same way every time β€” notifications, data entry, syncing, reports β€” runs automatically.

🧩 App integration

400+ pre-built integrations (plus a generic HTTP node) so any two services can exchange data without custom clients.

πŸ› οΈ Glue-code sprawl

Replace scattered cron scripts and Lambda one-offs with one visible, maintainable, centrally-run place.

Mechanics  How n8n works

A workflow is a graph of nodes. A trigger node starts it (a schedule, an incoming webhook, or an app event), then data flows node-to-node as JSON items, each node transforming or routing them until action nodes do something in the outside world.

A simple n8n workflow A schedule trigger runs an HTTP request, an IF node branches: if true, send a Slack message; if false, append a row to Google Sheets. true false ⏱ Schedule HTTP Request IF Send Slack Google Sheets

When a pre-built node doesn't exist, two escape hatches keep you unblocked: the HTTP Request node calls any REST API, and the Code node runs JavaScript (or Python) over the items β€” so a low-code workflow can drop into real code exactly where it needs to.

a Code node
// n8n "Code" node β€” runs over the items flowing through the workflow.
// Enrich each item, then hand them to the next node.
for (const item of $input.all()) {
  item.json.fullName    = `${item.json.first} ${item.json.last}`;
  item.json.processedAt = $now.toISO();     // n8n built-ins: $input, $now, $json…
}
return $input.all();

Character  What makes n8n distinct

🏠 Self-hostable

Run it on your own server (Docker/npm) so data never leaves your infrastructure β€” key for privacy, compliance and cost control.

πŸ§‘β€πŸ’» Code when needed

Low-code by default, but the Code and HTTP nodes mean you're never boxed in like a pure no-code tool.

πŸ€– AI / agent nodes

Built-in LangChain-based nodes let you build LLM chains and tool-using agents inside a workflow.

πŸ’Έ Predictable cost

Self-hosted is flat-cost rather than per-task β€” a big saver versus per-operation SaaS pricing at volume.

The trade-offs  Limitations

What to weigh before committing

License n8n is fair-code, not OSI open-source β€” distributed under the Sustainable Use License. Self-hosting for your own internal use is free, but offering n8n itself as a hosted service or embedding/white-labeling it commercially needs a licence. Read the terms if you plan to resell.

Ops Self-hosting has a cost of its own. Beyond a hobby instance you manage upgrades, a database (Postgres), and β€” to scale β€” "queue mode" with Redis and worker processes. That's real DevOps.

Scale Not a big-data engine. Items are processed largely in memory, so huge payloads or very high-throughput streaming can strain it. For heavy ETL or event streams, a dedicated pipeline (Airflow, Spark, Kafka) fits better.

Maintainability Large workflows get visually unwieldy; deep branching is hard to debug, testing is limited, and workflows are stored as JSON β€” versionable in git but not a first-class code-review/merge experience.

Integrations Some connectors are less deep or polished than a mature SaaS like Zapier's; you'll sometimes fall back to the raw HTTP node and wire auth yourself.

Net: n8n trades the turnkey polish (and cost) of hosted SaaS for flexibility, data control and code access. Great for technical teams; more setup than a point-and-click cloud tool.

Alternatives  Similar tools

The automation/integration landscape splits along a few axes: no-code vs code-first, cloud-only vs self-hostable, and app-glue vs data-orchestration.

ToolTypeHostingStyleNotes
n8nWorkflow automationSelf-host or cloudLow-code + codeFlexible, source-available, AI nodes
ZapierApp automationCloud onlyNo-codeBiggest app catalog, easiest; pricey at scale
Make (Integromat)App automationCloud onlyLow-code, visualPowerful visual builder, cheaper than Zapier
ActivepiecesWorkflow automationSelf-host or cloudLow-code + codeMIT-licensed n8n-style alternative
Node-REDFlow-based (events/IoT)Self-hostLow-code + JSOpen source; great for hardware/IoT/event flows
PipedreamDev automationCloud (mostly)Code-firstDeveloper-centric, code steps between apps
Workato Β· Tray.ioEnterprise iPaaSCloudLow-codeGovernance, SLAs, enterprise scale; $$$
Apache AirflowData orchestrationSelf-host / managedCode-first (Python)DAGs for data pipelines β€” not app-glue
TemporalDurable workflowsSelf-host / cloudCode-firstReliable long-running orchestration as code
Family resemblance: Zapier/Make/Activepieces are n8n's closest cousins (app automation). Node-RED shares the visual-flow idea for events/IoT. Airflow and Temporal solve a different problem β€” code-first orchestration for data and reliability engineers.

When to choose which

If you…Reach for…
Want self-hosting, data control, and code when neededn8n
Are non-technical and want the fastest, cloud-only setupZapier
Want a powerful visual builder that's cheaper than ZapierMake
Need a fully OSI-open-source, self-hosted optionActivepieces or Node-RED
Automate hardware / IoT / event streamsNode-RED
Build data pipelines / DAGs in PythonApache Airflow
Need mission-critical, long-running durable workflows in codeTemporal
Are an enterprise needing governance & SLAsWorkato / Tray.io
Rule of thumb. Non-technical + cloud β†’ Zapier/Make. Technical team wanting control, flexibility and cost efficiency β†’ n8n. Data engineering or reliability-critical orchestration β†’ Airflow/Temporal (different tool for a different job).

Cheat sheet

QuestionShort answer
What is n8n?A source-available, self-hostable workflow-automation tool (node-based).
What problem does it solve?Automating repetitive multi-app tasks without bespoke, hard-to-maintain glue code.
How is a workflow built?A trigger node feeds action/logic nodes; data flows as JSON items.
Escape hatch from no-code?The HTTP Request node (any API) and the Code node (JS/Python).
Biggest caveats?Fair-code licence, self-hosting ops, and not built for big-data scale.
Closest alternatives?Zapier & Make (cloud), Activepieces & Node-RED (open-source).

Knowledge check

Six questions on n8n. Pick an answer for each, then check your score.

1. What is n8n primarily for?

2. How is an n8n workflow structured?

3. When no pre-built integration exists, what keeps you unblocked?

4. Which is a real limitation of n8n?

5. A non-technical team wants the fastest cloud-only setup with the biggest app catalog. Best fit?

6. You need code-first DAGs for a data-engineering pipeline. Which tool is designed for that?

Next steps

πŸ”Œ APIs vs MCP

How tools connect to AI β€” n8n glues apps; MCP standardizes tools for agents.

🧰 Agent Development Kits

Build agents in code β€” the code-first counterpart to n8n's visual AI nodes.

πŸ“‘ IoT Γ— Agents

Event-driven automation β€” an agent triggered by a device, the AWS way.

πŸ““ All tutorials

The hub β€” LLM apps, embeddings, retrieval, agents, protocols, AWS & IoT.