
n8n bills by the execution, one run of a whole workflow, however many steps it has. On n8n's pricing page on 20 November 2026, n8n Cloud costs €20 a month billed annually for 2,500 executions on the Starter plan and €50 a month for 10,000 on Pro, or €24 and €60 billed monthly, and Starter allows 5 concurrent executions and 1 shared project. Business, from €667 a month billed annually for 40,000 executions, is sold for self-hosted n8n only, and Enterprise is priced by n8n's sales team. Every plan includes unlimited users and workflows and every integration.
Self-hosted, the Community edition has no license fee and no execution quota, because only paid plans have an execution quota. Its cost moves to the servers you run it on and the people who install, update and secure it, which n8n's own guide to choosing how to use n8n lists as your responsibility once you self-host. For a company whose data must stay in its own cloud, we recommend that route, with Business or Enterprise bought only when a feature the company needs comes with one of them.
| Community | Starter | Pro | Business | Enterprise | |
|---|---|---|---|---|---|
| Hosting | self-hosted | hosted by n8n | hosted by n8n | self-hosted n8n only | hosted by n8n or self-hosted |
| Price | no license fee | €20 a month billed annually or €24 billed monthly | €50 a month billed annually or €60 billed monthly | €667 a month billed annually or €800 billed monthly | no list price |
| Executions | no execution quota | 2,500 workflow executions a month | 10,000 or 50,000 executions | 40,000 to 500,000 executions a month | custom workflow executions |
| Adds | almost the complete feature set | 5 concurrent executions and 1 shared project | admin roles and workflow history | SSO and LDAP, multiple environments and version control using Git | audit logging, external secret store integration and log streaming |
n8n's pricing page lists Starter and Pro as hosted by n8n. Starter costs €20 a month billed annually or €24 billed monthly, for 2,500 workflow executions a month, with 5 concurrent executions and 1 shared project. Pro comes in two sizes, 10,000 or 50,000 executions a month: €50 a month billed annually or €60 billed monthly for the first, and €1,450 a year billed annually or €145 billed monthly for the second. It adds admin roles and workflow history, with up to 50 concurrent executions. A free trial of n8n Cloud gives you 1,000 executions and needs no credit card.
Business is for companies with fewer than 100 employees, and n8n sells it for self-hosted n8n only, since its pricing FAQ says the plan is not yet available on n8n Cloud. It costs €667 a month billed annually or €800 billed monthly at 40,000 executions, in sizes from 40,000 to 500,000 executions a month, and adds SSO and LDAP, multiple environments and version control using Git. Support comes through the community forum, and a company with fewer than 20 employees can apply for n8n's Startup plan, which takes 50% off Business.
Enterprise has no list price on n8n's pricing page, which lists custom workflow executions for it, and you can run it hosted by n8n or self-hosted. It adds audit logging, external secret store integration and log streaming, custom roles, 200+ concurrent executions and dedicated support with an SLA.
The Community edition is the free n8n you self-host, the edition that runs without a license key. It has almost the complete feature set, and n8n's comparison of editions lists what it leaves out, among them SSO, environments, external secrets, log streaming, projects, workflow sharing and version control using Git. Registering it with an email unlocks folders, debug in editor and custom execution data, and it stays free.
An execution is one run of a workflow from its trigger to its end. n8n's pricing page says of it: "It doesn't matter how many steps are in the workflow or how much data it processes. It's still a single execution." A 50-step enrichment workflow and a 3-step Slack alert each count once per run, so the number that moves your bill is how often your workflows run, and a trigger that fires when nothing needs doing spends executions without producing work.
Only production executions count toward a paid plan's quota, the runs that triggers, schedules and polling start on their own, according to n8n's page on understanding executions. A schedule counts every time it fires, a webhook counts every request that starts the workflow, and a polling trigger counts only when it finds new data. Test runs from the editor, sub-workflows that a parent workflow calls, error workflows, polls that find nothing and webhook requests rejected before the workflow starts do not count.
To estimate your count before you choose a plan, n8n's pricing FAQ suggests listing your workflows and how often each one runs, and notes that a daily schedule runs 30 or 31 times a month. Add the runs a day of every scheduled workflow to the events a day that start your other workflows, multiply by 31, the longest month, so that the count holds in every month, and pick the smallest plan whose quota covers it. One hourly workflow and 50 webhook requests a day come to 2,294 executions a month, inside Starter's 2,500.
Starter and Pro charge no overage fees, and the plan comparison on n8n's pricing page says "Starter and Pro are capped and have no overages." On Business, n8n charges €4,000 or $4,800 for each extra 300,000 executions, and its pricing FAQ says your workflows keep running past the quota. Enterprise overage fees are set in your order form. A single run can last up to 5 minutes on Starter and 40 minutes on Pro and Enterprise.
AI steps add a bill of their own. On n8n Cloud, a node that calls a model can use Gateway credits on Starter and Pro, or your own API key. Gateway credits are a prepaid balance that n8n draws down as nodes use it, and those nodes fail when the balance reaches zero, according to n8n's guide to Gateway credits. Gateway credits are not available on Enterprise or on self-hosted n8n, so there every model call goes on your own API key, billed by your model provider at its own rates and terms.
On n8n Cloud you pay for a Starter, Pro or Enterprise plan. Self-hosting takes n8n's fee off the bill, with no license fee on the Community edition, and puts the running costs on yours. In n8n's own comparison, n8n Cloud is fully hosted by n8n, and n8n handles hosting, updates and scaling. Self-hosted, you must provide and manage the infrastructure, maintenance is your responsibility, and technical expertise is required for installation and configuration. n8n's hosting guide names Docker Compose as the method for production deployments and has setup guides for AWS, Azure, Google Cloud, DigitalOcean and Hetzner.
Servers come from your cloud provider or your own hardware, at their prices. People install the instance, apply updates and security fixes, back up the database and watch for failed runs. At higher volume, queue mode adds Redis, which queues the executions, and workers that you add or remove as the load changes, according to n8n's page on queue mode.
| n8n Cloud | Self-hosted | |
|---|---|---|
| Hosting | fully hosted by n8n | you must provide and manage the infrastructure |
| Maintenance | n8n handles hosting, updates and scaling | maintenance is your responsibility |
| License | a Starter, Pro or Enterprise plan | no license fee on the Community edition |
| Data | in the EU, on servers in Frankfurt, Germany | wherever you host n8n |
| Model calls | Gateway credits on Starter and Pro, or your own API key | your own API key, billed by your model provider |
Business and Enterprise on your own servers still bill by execution. The license key that unlocks them must ping n8n's license server daily, and the ping includes the number of production executions. One key works on any number of instances, and their combined usage counts toward your quota. n8n's pricing FAQ gives the reason: "Our pricing is not tied to how much it cost us to run n8n but the value it brings to you and your organisation."
Where the data sits can settle the choice before the price does. n8n Cloud stores it in the EU, on servers in Frankfurt, Germany. Self-hosted, it sits wherever you host n8n, so workflows, credentials and execution history stay in your own cloud account.
The choice turns on how many executions you run a month and whether you have people to run servers. n8n's guide points to n8n Cloud when you want to start right away, have no technical expertise or don't want to manage infrastructure, and to self-hosting when you need full control over deployment and have the infrastructure and technical resources to run it. Workflows export as JSON, so a team can start on n8n Cloud and move later.
On your own servers, start on the Community edition and buy a license only for a feature you need. Business is the step up when sign-in must go through your identity provider with SSO and LDAP, or changes must move through environments and Git before they go live. Enterprise is the step up when you need audit logging, log streaming to external services or an external secret store, or dedicated support with an SLA.
We deploy n8n Business or Enterprise in your own cloud account and choose the plan with you before you talk to n8n sales. We first estimate how many executions you expect and list your security requirements, so you buy Business or Enterprise only when you need it. On an instance you already run, we look for repeated API calls, model calls on records that have not changed and heavy workflows that should run in queue mode, since each one shows up on the bill or as a slow run.
On n8n Cloud, Starter has the lowest price, at €20 a month billed annually or €24 billed monthly for 2,500 executions on n8n's pricing page on 20 November 2026, and Pro costs €50 or €60 a month for 10,000. Self-hosted, the Community edition has no license fee and no execution quota, and you pay for the servers and the people who run them instead.
An execution is one run of a whole workflow, from its trigger to its end, with any number of steps and any amount of data. Only production executions count toward a paid plan's quota, the runs that triggers, schedules and polling start on their own. Test runs from the editor, sub-workflows, error workflows and polls that find no new data do not count.
The Community edition is free to self-host, with no license fee and no execution quota, and it has almost the complete feature set. You still pay for the servers it runs on and for the time of the people who install, update and secure it. Business and Enterprise on your own servers need a paid license key and bill by execution.
Starter and Pro charge no overage fees, since n8n's pricing page lists both as capped with no overages. On Business, n8n charges €4,000 or $4,800 for each extra 300,000 executions and your workflows keep running, and Enterprise overage fees are set in your order form. Moving to a larger size of Pro or Business raises the quota.
Not yet. n8n's pricing FAQ says the Business plan is only available for self-hosted n8n, with a waitlist for a version hosted by n8n. On n8n Cloud, Pro is the largest plan with a list price, at up to 50,000 executions a month, and Enterprise covers more on terms set with n8n's sales team.
No, model calls are a bill of their own, apart from the plan's executions. On n8n Cloud Starter and Pro, a node that calls a model can draw on Gateway credits, a prepaid balance that n8n draws down as nodes use it, or use your own API key. On Enterprise and on self-hosted n8n, every model call goes on your own API key, and your model provider bills it at its own rates and terms.
Choosing an n8n plan? Book a discovery call, and a senior partner estimates your executions and security requirements with you and recommends a plan before you talk to n8n sales.