Make the hosting decision around one workflow
Start with the automation you want to keep working. An illustrative SaaS intake flow receives a form, checks required fields, creates a CRM record, and notifies the right person. List the systems it touches, how often it runs, and the consequences of a missed or repeated execution.
Then ask what actually requires infrastructure control. An existing network boundary or deployment policy is a concrete requirement. General discomfort about a subscription is a different question: it calls for a cost comparison that includes the time and responsibility of operating the replacement.
How Cloud pricing differs from a server bill
n8n's current pricing page describes paid plans in terms of monthly workflow executions, rather than charging for each step. Hosted plans also have execution and collaboration limits that matter to a production workflow. Check the selected billing period, concurrency, and required plan features together.
Estimate scheduled checks as well as customer-triggered runs. A workflow polling every few minutes can execute many times even on quiet days. Keep model API charges and other connected-service fees as separate budget items. Paying n8n does not establish that every external service a workflow calls is included.
Self-hosted editions have different capabilities
n8n can run on your own infrastructure. Without a paid license it runs as Community edition; Business and Enterprise licenses enable additional capabilities. The official edition comparison lists features absent from Community, including projects, workflow and credential sharing, SSO, and Git-based version control.
That distinction matters when a workflow stops belonging to one person. If an operations colleague must edit it or use shared credentials, verify the actual edition's collaboration model before calling a small server an equivalent replacement for Cloud. Registration unlocks some Community extras, but it is not the same thing as buying all paid features.
Compare ownership as well as features
The hosting documentation covers deployment and configuration on your own infrastructure. In your evaluation, assign an owner for upgrades, persistent storage, credential protection, monitoring, and recovery. These are continuing tasks, even when the first installation is quick.
For the intake example, write a recovery exercise: the instance becomes unavailable after a form is received but before the CRM record is created. Determine how you will detect the incomplete work, recover the instance, and resume without creating a duplicate contact. A backup file is useful only if the team can restore the required data and configuration.
| Decision | Cloud evaluation | Self-hosted evaluation |
|---|---|---|
| Operations | Understand managed-service boundaries | Assign an infrastructure and recovery owner |
| Capacity | Check execution and concurrency allowances | Size and monitor the infrastructure |
| Team access | Check the chosen hosted plan | Check Community versus paid-edition features |
| Data paths | Review where connected data travels | Review the instance and every external integration |
Put a value on maintenance time
Use a simple monthly comparison: infrastructure plus required license plus external services plus maintenance time. For an illustrative calculation, a server at €15 and two hours valued at €40 per hour total €95 before other services or incident work. Those are example assumptions, not n8n prices or a forecast of your maintenance burden.
Cloud also consumes time: you still design workflows, maintain integration credentials, investigate failures, and monitor volume. Compare the work that changes between options instead of assigning all engineering time to one side. A team already operating similar infrastructure may reach a different decision from a solo founder with no on-call capacity.
Pilot reliability before moving critical work
Try one representative flow with synthetic records. Check a successful run, an expired credential, an unavailable downstream API, a repeated input, and recovery after an interruption. Record both the workflow outcome and the effort required to diagnose it. These observations are more useful than comparing how quickly each option produces its first successful run.
Choose Cloud when its current plan and managed environment satisfy your needs at an acceptable total cost. Choose self-hosting when the control it provides is valuable enough to justify an explicitly owned operating plan. Keep the existing process available until you can explain how failed and duplicate work will be handled.
Sources and editorial method
Official documentation linked in this guide was checked on October 26, 2026. Product behavior, pricing, and plan limits can change. The workflow examples are editorial proposals, not reports of a hands-on benchmark or customer results. Confirm the current documentation before implementing the workflow.