Loading tutorials…
Loading tutorials…
Self-hosted n8n is great if you like managing Postgres. n8n Cloud is great if you like building workflows. This walks through the Cloud setup path — plan choice, environments, users, first workflow — in under an hour.
Who this is forOperators choosing n8n Cloud over self-hosting because they want the workflow power of n8n without owning infrastructure. Also for teams evaluating n8n against Zapier or Make on the same surface (managed SaaS, no DevOps).
What you'll need
Step 1
n8n.io → Sign up for Cloud. Pick the closest data region. The 14-day trial gives full Pro features.
Go to n8n.io → "Get started for free." Use a real business email; verification is required.
During signup, you pick a data region: US, EU, or APAC. This cannot be changed later without a full export/import. Pick the region where most of your data already lives.
You land in the Cloud editor with a 14-day free trial of the Pro plan. Use the trial to exercise Pro-tier features (AI Agent, advanced executions, multiple users) so you know what you are paying for.
You will not be charged until you actively pick a plan at the end of trial. There is no auto-upgrade.
Step 2
Starter (~$24/mo), Pro (~$60/mo), Enterprise (custom). Plans are gated by executions/month, active workflows, and feature access — not seats.
Starter: ~$24/mo for 5,000 executions, 10 active workflows, 1 shared project. Right for solopreneurs and tiny teams testing.
Pro: ~$60/mo for 10,000 executions, 50 active workflows, 5 projects, AI Agent node, SSO. Right for most growth teams.
Enterprise: custom pricing — unlimited executions, audit logs, environments, log streaming, dedicated support. Required for regulated industries or 100+ workflows.
Critical: one execution = one full workflow run, not one node. A 30-node workflow that runs once = 1 execution. This makes n8n significantly cheaper than Zapier (per-Task) for complex workflows.
Right-size based on YOUR expected monthly volume, not trial volume. Track executions in Settings → Usage during the trial.
Step 3
Settings → Users. Add teammates with roles: Owner, Admin, Editor, Viewer. Use SSO if you are on Pro+.
Open Settings → Users. Click "Invite users."
Roles: Owner (1, billing access), Admin (can manage users + workflows), Editor (build and edit), Viewer (read-only).
For each invitee, enter the email and pick a role. They receive a setup link that expires in 7 days.
On Pro+, enable SSO via Settings → SSO. Supports Google, Microsoft, and SAML. Once SSO is on, password login is disabled for org members.
For agencies and contractors: use Viewer or Editor (not Admin) so they cannot manage users or billing.
Step 4
Create a "Production" and a "Staging" environment so you can build and test changes safely.
On Pro+: Settings → Environments. By default, you have one environment. Add a second called "Staging" or "Dev."
Each environment has its own set of credentials and workflows. Promote workflows from Staging to Production via export/import.
Rule of thumb: never edit a Production workflow directly. Always edit in Staging, test, promote.
On Starter (single environment): use clear naming — prefix experimental workflows with `[DRAFT]` or `[TEST]` so you do not accidentally activate a half-built workflow against production data.
Step 5
Create a Schedule Trigger → HTTP Request workflow to validate the Cloud install actually executes.
In the editor: New Workflow.
Add a Schedule Trigger node. Set it to "Every Day at 9:00 AM" or "Every 5 minutes" for faster testing.
Add an HTTP Request node. Method GET, URL `https://api.github.com/repos/n8n-io/n8n`. No auth needed.
Save and activate the workflow.
Wait for the next scheduled run (or click Execute Workflow manually to trigger immediately). Confirm the execution appears under Executions → Successful with the GitHub API response in the node output.
Step 6
Settings → Usage shows current execution count. Settings → Billing has alerts. Set an alert at 80% of plan limit so you can upgrade before hitting the cap.
Settings → Usage shows execution count, active workflow count, and remaining quota for the month.
In Billing, set an email alert at 80% execution usage. Cloud does NOT halt workflows at the limit — but Enterprise plans gate certain features.
For Starter and Pro, if you exceed your monthly execution count, n8n charges overage at a clear per-execution rate. Watch the dashboard to avoid surprise bills.
Schedule a monthly Usage review on the calendar. If you trend toward the limit two months in a row, upgrade BEFORE the third overage.
Common mistakes
Picking the wrong data region
What goes wrong: You sign up in US during trial because that is the default. Three months in, an EU customer asks for proof of data residency in the EU. You have to export every workflow, recreate the account in EU, reconnect every OAuth credential. Two days of work and a customer concession on launch date.
How to avoid: Pick the region matching your customer base or regulatory requirements at signup. EU for GDPR-sensitive data, US for North American business, APAC for Australia/SEA-heavy operations.
Treating "active workflows" as a soft limit
What goes wrong: Starter is capped at 10 active workflows. You build the 11th, hit the cap, and rather than upgrading you toggle off an existing workflow 'temporarily' — and forget. That workflow misses a week of triggers. By the time someone notices the missing data, recovery is impossible.
How to avoid: Track active workflow count alongside executions. When you cross 8 of 10 active, plan the upgrade. Do not toggle off production workflows to make room.
Sharing one Owner login across the team
What goes wrong: Two ops folks share one login. Both edit the same workflow at the same time. One overwrites the other's changes — no version history, no recovery. Six hours of work lost.
How to avoid: Every teammate gets their own user account. Even on Starter, invite at least one other Admin so the Owner is not a single point of failure.
Skipping environments and editing in production
What goes wrong: You edit a live workflow during business hours. A field-mapping mistake causes 200 records to write with wrong values before you notice. Cleanup takes 4 hours of database work plus an apology email to affected customers.
How to avoid: On Pro+, use Environments. On Starter, duplicate the workflow, edit the copy, test, then deactivate the old and activate the new. Never edit live production workflows.
Storing credentials in workflow expressions instead of Credentials
What goes wrong: You hardcode an API key into an HTTP Request URL inside a workflow. Six months later, that workflow is duplicated three times and the key is now in four places. Rotating the key requires finding all four — and you miss the third.
How to avoid: Always use Credentials (Settings → Credentials). One stored credential gets referenced by name from any workflow. Rotating means updating one record.
Recap
Done — what's next
How to set up n8n self-hosted
Read the next tutorial
Hand it off
n8n Cloud removes the infrastructure problem; it does not remove the build problem. Most teams that adopt Cloud still end up working with an automation specialist on workflow design, error handling, and the ongoing maintenance of 20-50 active workflows. EverestX matches at $14-16/hr, typically $400-1,000/mo for ongoing work.
See specialist rates
For complex multi-step workflows, yes — usually 50-70% cheaper. n8n bills per workflow execution; Zapier bills per node/Task. A 20-node n8n workflow = 1 execution; the equivalent Zapier Zap = 20 Tasks. For simple single-step automations, the gap narrows.
Yes. Export workflows as JSON, set up self-hosted, import workflows, reconnect credentials. Plan a 1-2 day migration window for 20-30 workflows. Note: execution history does not migrate — start fresh.
One execution = one complete workflow run from trigger to final node. A workflow with 50 nodes that runs 100 times in a month = 100 executions. This makes complex workflows much cheaper on n8n than on Zapier.
Yes, on Pro and Enterprise plans. AI Agent uses LangChain under the hood and supports OpenAI, Anthropic, Mistral, and local models. Starter does not include AI Agent — you would need to upgrade.
On Cloud, workflows do not halt at the limit. You are billed overage at a clear per-execution rate. Set the 80% alert and upgrade proactively to avoid surprises.
n8n
Self-hosting n8n is cheap until the day you upgrade and lose every workflow because the docs let you start on SQLite. This walks through the durable install path — Docker + Postgres + HTTPS — that survives version bumps and restarts.
n8n
You have n8n running. Now build one real workflow that does something useful. This walks a real trigger → action chain end-to-end with the expression-mapping detail most tutorials skip.
n8n
Most 'AI Agent' tutorials build a demo that breaks the second real users touch it. This walks the n8n Agent node end-to-end — model selection, memory, tool wiring, and the guardrails that make it production-ready.
n8n
DIY n8n is great until you have 15 workflows and a credentials audit you keep deferring. This is the honest framework: when the cost of self-managing exceeds the cost of a specialist, and how to tell which side you are on.
Zapier
You signed up for Zapier and the dashboard is staring at you. This walks through one real, working Zap end-to-end — trigger app, action app, sample data, test, turn on — without the marketing fluff.