Choose between n8n Cloud and self-hosting by deciding who will operate the system. The server bill is only one part of the decision. Someone still needs to own workflow correctness, credentials, failure handling and the records the workflow changes.
For a small team, n8n Cloud is often a practical starting point when nobody is assigned to maintain infrastructure. Self-hosting can fit a defined deployment or customization need when a capable owner has time to operate it. This is Qixazow’s decision framework, not a claim that one option is always cheaper or more secure.
Separate deployment from the plan you need
n8n’s deployment choice documentation makes two decisions explicit: Cloud versus your own infrastructure, then the plan or edition that provides the required features and limits. Both deployment options support production use.
Community edition, paid self-hosted plans and Cloud plans have different capabilities. Check the current official plan comparison for collaboration, execution limits, history, governance and support. A deployment choice alone does not establish which of those features you receive.
Compare responsibilities for your workflow
| Decision | n8n Cloud | Self-hosted n8n |
|---|---|---|
| Infrastructure owner | n8n manages the hosted platform. | Your team or contracted operator manages the deployment. |
| Environment control | You work within the hosted configuration and plan options. | You control the environment and must maintain that configuration. |
| Workflow responsibility | Your team owns the inputs, logic, access, external actions and acceptance checks. | The same workflow responsibilities apply, alongside infrastructure operation. |
| Cost to evaluate | The subscription, usage, connected tools and workflow delivery effort. | Hosting, any required license, backups, monitoring and the operator’s time, plus connected tools and workflow effort. |
| Recovery planning | Check the plan’s history and retention, and keep a workflow-level recovery runbook. | Also own recoverable infrastructure backups and a tested restore process. |
The infrastructure distinction is documented by n8n. The workflow, cost and recovery responsibilities in this table are our operating checklist.
Use the small-team test
Before deciding, answer these questions for one representative workflow:
- Who is the operator? Name the person who handles upgrades, incidents and access changes. If that role is unfilled, price the missing operational work instead of assuming it is free.
- What must the environment support? List private-system access, custom dependencies and deployment controls. Verify the required capability in the official documentation rather than assuming a feature is available everywhere.
- What data enters the workflow? Map source systems, execution data, model providers and destinations. Choosing self-hosting does not stop a workflow from sending data to an external API.
- What does ordinary load look like? Use representative inputs, file sizes, execution frequency and overlapping runs. Do not choose capacity using a tiny editor test.
- How long is evidence needed? Decide what execution data supports troubleshooting, who can read it and when it should be removed.
- What happens when it fails? Define a manual fallback, alert owner, recovery path and acceptance checks.
Self-hosting includes operational work
The official hosting documentation covers installation, configuration, architecture and keeping the instance running. Treat a working installation as the beginning of operation, rather than proof that backup, recovery and maintenance are complete.
For credential recovery, preserve the required encryption key securely along with a documented restore process. n8n explains that credentials stored in its database are encrypted with that key in its encryption key documentation. A protected backup must stay outside public access and be restorable by the authorized operator.
n8n’s update guidance recommends checking release notes and testing changes before applying them. Choose a test method that your plan and infrastructure support. The operator also needs access controls, secure connections and monitoring appropriate to the instance.
Cloud still needs workflow care
A managed platform removes infrastructure tasks; it does not choose your acceptance criteria or reconcile a duplicated customer reply. n8n’s Cloud data management guide notes that inefficient workflows can exceed memory or storage limits. Check current plan capacity and save only the execution data you actually need.
Regardless of hosting, validate important inputs, route exceptions to an owner, and verify writes in their destination. If tickets are missing or acknowledgements repeat, our n8n support workflow repair guide starts with the affected request and completed actions.
Make the choice around one scoped process
Do not migrate a working system solely because another deployment appears cheaper. Establish its constraints, actual operating effort, required capabilities and recovery plan first. Compare the total responsibility that comes with each option.
Our Operations Audit maps one process and its ownership, dependencies and delivery scope. Discuss an n8n assessment if you need a decision grounded in your team’s system. Qixazow quotes the work after discovery and is independent of n8n.
Sources and review date
Official n8n pages linked above were checked on 14 October 2026. Pricing, plan limits, interface labels and deployment features change over time; verify them before purchasing or migrating. The small-team test and total-cost framework are Qixazow’s recommendations, not a security certification or promised saving.
Prepared with AI assistance using the sources listed in this guide. Examples and operating checklists are Qixazow guidance. Testing and measured outcomes are stated only where evidence is available.