n8n Pricing: Cloud Plans and the Cost of Self-Hosting

We checked the prices and terms in this article against the n8n pricing page and official docs on October 2, 2026. Prices are shown in euros. Prices and plans may change. Before you sign up, check the latest information in the sources at the end.

How n8n pricing works depends on how you use it. With n8n Cloud, run by n8n, you pay a plan fee. With self-hosting on your own server, instead of paying n8n, you pay for the server and the operating effort.

n8n Pricing Is Based on Executions

According to the pricing page, n8n pricing is based on the number of workflow executions per month. An execution is one run of an entire workflow from start to finish. It counts as one no matter how many steps the workflow has or how much data it handles.

All plans have no limit on users or workflows. You can use all integrations (nodes).

Estimating how many executions you need

The pricing page shows how to estimate by use case.

• Scheduled runs: count how many times it runs per month. A month has 30–31 days, so once a day means 30–31 runs. Every 5 minutes means about 8,600–8,900 runs
• Webhook runs: the number of triggering events per day equals the number of executions per day
• Chatbots: multiply the number of conversations per week by the average number of messages per conversation

You can check your usage on the Insights dashboard.

Cloud Plan Differences

n8n Cloud plans are Starter, Pro, and Enterprise. Business is currently available only for self-hosting. Paying annually is 17% cheaper, according to the pricing page.

Starter vs. Pro

• Starter: 1 shared project, up to 5 concurrent executions, up to 5 minutes per execution, 7 days of log retention
• Pro: 3 shared projects, up to 50 concurrent executions, up to 40 minutes per execution, 30 days of log retention. Includes admin roles and workflow history
• Both: AI agents, MCP, and AI nodes. Monthly credits for n8n Assistant (1,600 on Starter, up to 9,600 on Pro)

Limits on saved executions

n8n Cloud saves records of recent executions for review and troubleshooting. Starter saves up to 2,500 records and Pro up to 25,000. Storage is 2.5GB on Starter and 25GB on Pro. When any limit on count, storage, or retention days is reached, the oldest records are deleted. Reaching the limit does not stop workflow executions.

n8n Cloud data is stored on servers in Frankfurt, Germany, within the EU. With self-hosting, data is stored wherever you choose.

Business and Enterprise

Business is a plan for companies with fewer than 100 employees. It adds SSO (SAML, LDAP), multiple environments, queue mode, and version control with Git. After you sign up, a license key arrives by email, and you activate it in your own n8n. Companies with fewer than 20 employees that meet the conditions can use the Start-up Plan, which gives 50% off Business.

Enterprise sets execution volumes individually and requires contacting sales. On Business and Enterprise, workflows do not stop when you exceed the limit. However, overage charges may apply if you do not move up to the next tier. The Business overage charge is €4,000 per 300,000 executions.

What You Can Use Without Paying

When we checked the pricing page on October 2, 2026, n8n Cloud plans started with the paid Starter plan. No plan without ongoing charges was shown. Instead, there is a 14-day trial.

• The trial lasts 14 days and lets you try Pro features
• Up to 1,000 executions and up to 5 concurrent executions
• The n8n Cloud trial does not require a credit card. The Business trial does
• If you do not upgrade after the trial, the workspace is deleted. You can export workflows for up to 90 days after the trial ends

To keep using n8n at no charge, use the self-hosted Community edition. The official docs describe it as an edition that includes almost all features and can be used without a time limit. Registering an email address adds features such as folders. It does not include SSO or sharing of workflows and credentials.

Self-Hosting Costs: Servers and Operating Effort

With the Community edition, you pay no license fee to n8n. Instead, you cover the following costs and work yourself.

• Server costs: a cloud provider or your own servers. The official guide for building a Docker Compose setup by hand calls for at least 4GB of memory and at least 2 vCPUs
• Domain and certificates: needed for external access
• Operating effort: updates, backups, incident response, and security settings

The official docs say n8n performance depends on the type of workflows, available resources, and scaling settings. If you hit errors from insufficient memory, self-hosting means adding server memory, which can increase what you pay your hosting provider. We explain setup in n8n self-hosting.

Using a paid plan with self-hosting

If you need SSO or multiple environments, add a Business or Enterprise license key. In that case, the license key contacts the license server daily and sends usage data such as production executions. One license key can be used on any number of instances. Executions across all instances count toward the limit.

AI API Fees Are Charged Separately

When you build AI agents, AI model usage fees apply in addition to your n8n plan. There are two ways to pay.

• Use your own API key: contract directly with Anthropic, Google, and others, and get billed by each. Each provider prices by tokens. Anthropic and Google's Gemini API list prices per 1 million tokens
• Use Gateway credits: available on n8n Cloud Starter and Pro. Each request is deducted from a prepaid balance at the rates in n8n's service price list

Gateway credits are not available for self-hosting or n8n Cloud Enterprise. When the balance reaches zero, nodes that use Gateway credits stop. Nodes that use your own API key are not affected. Purchased top-up credits expire 12 months after purchase.

With Gateway credits, n8n sends requests to the model provider from its own account. The docs say the provider receives the request content, such as prompts, but not the user's identity or n8n account information. If you handle customer data, design on the assumption that data goes to the model provider with either payment method.

Credits for n8n Assistant are a separate balance from Gateway credits. They cannot pay for workflow nodes. We also cover the full picture of AI agent costs in AI agent costs.

Choosing the Right Option

If you are unsure, start with the 14-day trial and see how many executions you actually run. Then choose a plan that fits that number, or decide to move to self-hosting.

If you sell the workflows you build

Self-hosted n8n is provided under the Sustainable Use License. The license page allows charging for consulting and support, such as building workflows. We explain the commercial terms in detail in n8n commercial use.

On Employee Store, you can list AI agents, including n8n workflows. There is no listing fee or upfront cost. The fee is 20% of the deal amount and applies only when a deal closes. For details, see the seller guide.