Is self-hosted n8n cheaper than n8n Cloud?
Self-hosted n8n is cheaper at scale and more expensive in attention. Cloud plans bundle hosting, upgrades, and backups into a predictable monthly fee, while self-hosting trades that fee for a server bill plus the hours you spend patching, monitoring, and restoring. Below roughly 20 active workflows the difference is small; above that, self-hosting usually wins on cost — provided someone owns the operations. Count your time at a real rate before deciding.
What does n8n Cloud actually include?
Cloud covers managed hosting, version upgrades, backups, and support, and removes the need for a server administrator. That is genuinely valuable for a team without infrastructure skills, and it shortens time to first workflow to hours. The trade-offs are execution or tier limits, less control over the runtime, and data residency determined by the provider. For regulated clients or strict procurement, those constraints can decide the question before price does.
What do you actually own when you self-host?
You own the deployment: the server, the database, credentials, workflow JSON, and execution history. That matters for client work, because workflows encode business logic and are the asset a client is paying for. An owned install lets you hand over the whole system, move providers, or run it inside a client's cloud account. Ownership also means you inherit patching and backup duty — the benefit and the bill arrive together.
What are the hidden operating costs?
Budget for server and database hosting, backup storage, monitoring, TLS certificates, upgrades, and incident response. A single-node install is modest; queue mode with workers, Redis, and a separate database costs more and needs tuning. The largest line is rarely infrastructure — it is the person who notices when executions stop. Without an alert on failed workflows, a silent outage can run for days and cost more than a year of cloud fees.
When is n8n Cloud the better choice?
Choose cloud when automation is supporting rather than core, when nobody wants to own infrastructure, or when you need a vendor to call during an incident. It is also the sensible default for a first proof of concept, because it removes setup risk while you validate whether the workflow is worth keeping. Migrating to self-hosted later is straightforward if you export workflows and document credentials. Decide on operations capacity, not sticker price.
How do you migrate without downtime?
Export workflows and credentials, stand up the target instance, restore the database, then replay a sample of recent executions to confirm behaviour. Run both systems in parallel on non-critical workflows for a week before cutting over. Keep the old instance read-only for a month, and document every credential and webhook URL — webhooks are the usual cause of a silent post-migration failure. Test failure alerts before you trust the new host.
Frequently asked questions
Do I need a DevOps engineer to self-host n8n?
Not for a single-node install, but you need someone comfortable with Docker, DNS, TLS, and backups. If nobody on the team can debug a failed container at 9am, cloud is cheaper than learning on a client's production system.
Is self-hosted n8n secure enough for client data?
It can be, if you apply the basics: patching, least-privilege credentials, network restrictions, encrypted backups, and audit logging. Security is a configuration outcome, not a hosting model — cloud instances are also misconfigured regularly.
Can I white-label n8n for clients?
Yes, subject to the licence terms of the version you use. Check the current licence before reselling, and confirm whether your use case requires a commercial agreement. Ownership of the deployment does not automatically grant resale rights.
Want the full playbook?
The free 2026 AI Revenue Playbook turns these shifts into a plan with costs, scripts and pricing templates.