What n8n does
n8n is a workflow automation platform combining visual nodes, code, APIs, data transformations, and AI integrations with hosted or self-managed deployment.
n8n occupies useful middle ground between no-code automation and a custom integration service. Builders arrange nodes visually, but can drop into expressions, JavaScript or Python, direct HTTP requests, webhooks, and custom nodes when a packaged connector is insufficient. That combination suits technical operations and engineering-adjacent teams that want faster iteration without surrendering control over every transformation. AI-oriented nodes can connect models, memory, tools, vector stores, and data sources, although reliability still depends on the workflow around the model.
Cloud pricing is based primarily on production executions: one run of an entire workflow counts as one execution even when it contains many steps or processes significant data. That can make a complex workflow easier to forecast than per-step billing, but loops, sub-workflows, chat messages, concurrency, retention, and plan limits still shape cost. n8n's live pricing calculator varies by currency, cadence, and execution tier, so buyers should verify the current quote. The official page describes a 14-day trial and a Starter allowance of 2,500 executions, while its AI Assistant remains an early preview with separate monthly credits on Cloud plans.
Deployment is the defining choice. n8n Cloud reduces infrastructure work and currently stores hosted-plan data in Frankfurt, Germany. Self-hosting can place data and runtime in an environment the organization controls and unlock operational flexibility, but it does not make the installation secure automatically. The operator owns patching, encryption, backups, secrets, network exposure, scaling, monitoring, and recovery. Teams without those skills should not choose self-hosting merely because the software can be deployed that way.
For AI automation, use a deterministic shell around probabilistic steps. Validate schemas, constrain tools, redact unnecessary data, cap loops, and add approval before messages, deletions, payments, or sensitive updates. Save enough execution evidence to debug failures without retaining confidential payloads longer than necessary. n8n is less attractive for a nontechnical department that wants vendor-managed connectors with minimal maintenance, or for a tiny two-app automation already available natively. It excels when custom APIs, branching, code, hosting choice, and end-to-end execution economics materially improve the process.
How n8n works
An n8n workflow begins with a trigger and passes structured items between nodes. Nodes call applications, APIs, databases, code, or model providers; expressions map prior outputs into later inputs, while branches and error workflows control exceptional paths. n8n Cloud counts a complete production workflow run as an execution rather than billing every node separately.
Begin one complete execution
A webhook, app event, schedule, chat message, queue item, or manual action creates a workflow execution. Early nodes validate authentication, schema, idempotency keys, and required data before downstream systems or models are called.
Transform and enrich items
Nodes read applications, APIs, databases, code, or AI providers and pass structured items between steps. Expressions map earlier output into later parameters; branches, loops, merges, and sub-workflows encode the process while Cloud billing generally counts the full run as one execution.
Validate uncertain output
Model and agent nodes should receive narrow tools, explicit schemas, bounded iteration, and only necessary context. Deterministic nodes check classifications, citations, formats, and policy; approval pauses protect external messages and irreversible updates.
Operate the deployed workflow
Execution records and error workflows support debugging, retries, and alerts. Teams must tune retention, concurrency, timeouts, credentials, and provider limits; self-hosted operators additionally own upgrades, networking, backups, scaling, encryption, and disaster recovery.
How to set up n8n
Choose Cloud or self-hosted
Assess data location, engineering capacity, uptime, patching, backup, and support requirements before selecting the deployment model.
Model one production workflow
Define its trigger, input schema, expected result, monthly run volume, owner, service accounts, and recovery objective.
Configure credentials safely
Use least-privilege credentials and the platform's credential store; keep secrets out of node parameters, code, sample data, and logs.
Build validation and recovery
Map fields explicitly, validate model output, handle rate limits and empty data, and route failures to an owned alert or error workflow.
Test and operate
Exercise duplicates, retries, malformed inputs, permission failures, and provider outages before activation; then monitor executions, concurrency, cost, and retention.
n8n FAQs
Is n8n free?
n8n offers a self-hosted Community Edition and a time-limited Cloud trial. Self-hosting still carries infrastructure and operations costs, and some organizational features require paid licenses.
How does n8n Cloud count usage?
Its main unit is a production workflow execution: one complete run counts once regardless of the number of ordinary steps, subject to current plan definitions and separate limits.
Where is n8n Cloud data hosted?
n8n's current pricing FAQ says hosted-plan data is stored on servers in Frankfurt, Germany. Verify contractual and regional requirements before deployment.
Does self-hosting make n8n private by default?
No. It gives the operator control, but that operator must correctly configure access, networking, encryption, secrets, backups, updates, logging, and retention.
Listing reviewed 2026-07-15. Product details and pricing can change; verify important terms on the provider's website.
Related Automation AI tools
Related AI guides
Reviews
Tell the community what you made, what worked, and what you wish you knew before starting.