n8n Cost Calculator: AI Agent Cost per Month

This free n8n cost calculator shows your likely monthly bill. Monthly plan priced by workflow executions. AI model cost is extra. Your real cost depends on your own usage, so enter your numbers above. We use the official n8n pricing page, last checked on November 3, 2026.

What is n8n?

n8n (say "n-eight-n") is a workflow automation tool. You build automations by connecting blocks, called nodes, on a canvas. One node might catch a new form entry, the next asks an AI model to write a reply, and the next sends the email and updates your CRM. n8n has AI Agent nodes, so a workflow can use a model to decide which tool to call next.

n8n is popular with agencies and technical teams because it is flexible and you can run it on your own server for free. You can also let n8n host it for you on their cloud. Either way, you still pay for the AI model your agent uses, so an n8n bill has two parts: the n8n plan and the model tokens.

How much does n8n cost?

n8n charges by workflow executions: one full run of a workflow from start to finish, no matter how many steps it has. That is different from tools that charge per step or per task. A workflow with 30 nodes still counts as one execution. Every plan includes unlimited users, unlimited workflows and every integration.

These are monthly billing prices. If you pay yearly, n8n says you save about 17%: Starter drops to $20 a month, Pro to $50 and Business to $800. Companies with fewer than 20 employees may also qualify for a Startup plan with 50% off Business.

Here is what one execution costs if you use your full allowance: about $0.0096 on Starter, $0.0060 on Pro and $0.0240 on Business. So the per-run cost drops on Pro, but Business costs more per run. You pay for Business for features like SSO, Git version control and queue mode, not for cheaper runs.

Limits that matter for AI agents

• Run time: a single execution can run up to 5 minutes on Starter and 40 minutes on Pro. Long research agents may need Pro.
• Concurrency: Starter runs 5 executions at the same time. Pro allows up to 50. Busy chat or webhook agents can hit the Starter limit at peak times.
• Saved history: Starter keeps up to 2,500 saved executions; Pro keeps 25,000. This matters when you need to debug a bad answer from last week.

What is included, and what is not?

Included in your n8n plan

• The visual editor, all integrations (nodes), and the AI Agent, MCP and AI nodes.
• Hosting, if you use n8n Cloud (Starter and Pro).
• n8n Assistant credits for help building workflows: 1,600 a month on Starter and up to 9,600 on Pro. These help you build, they do not run your agent.
• Code steps in JavaScript or Python, webhooks and API access.

Not included

• AI model tokens (you connect your own OpenAI, Claude or other key)
• Server cost if you self-host
• Phone and voice minutes
• Email or dedicated support on Starter and Pro (forum support only).

n8n Cloud plans also list "AI models without API keys" and gateway credits, which let you try models without your own key. For a production agent, most teams connect their own OpenAI, Claude or other key, so our calculator adds the model cost at the provider's list price.

What does n8n cost at 3 business sizes?

This example is a lead follow-up agent: each new lead runs one workflow that writes a personal email with GPT-6.1 Terra and logs it. That is one execution per lead. We use 22 working days a month and no safety buffer.

At the small size, the n8n plan is 89% of the bill. The AI model costs just $3.01, and switching to GPT-6 Luna would cut the total only to $24.31. For comparison, the same agent coded directly on the OpenAI API costs about $3.01 in tokens, but you would need a developer and a server.

At the medium size you need 6,600 executions, which fits Pro. At the large size you need 33,000 executions, more than Pro's 10,000, so the calculator moves you to Business, which is self-hosted. Before you upgrade, check whether batching makes sense: one execution that handles 10 leads at once counts as one run, not 10.

A heavier research agent tells a different story. At 20 reports a day on GPT-6.1 Terra, the model costs $35.06 a month and the n8n plan is only $24. For heavy agents, the model, not n8n, drives the bill.

What hidden costs should you watch for?

• Self-hosting is free, the server is not. The Community Edition costs $0, but you pay for the server, backups and your time to update and secure it. Business is also self-hosted, so the $960 license is not the full cost.
• Test runs. Check how test runs in the editor are counted for your plan before you load-test a workflow.
• Polling triggers. A workflow that checks for new emails every minute can start many executions, even when there is nothing to do. Use webhooks where you can.
• Going over your plan. n8n says overage fees on Starter and Pro are capped, and Enterprise is custom. Watch your usage page so a busy month does not surprise you.
• Error workflows and retries. Each retry and each error workflow can be another execution, and another set of model tokens.

How can you lower your n8n bill?

1. Batch your work. Process items in groups inside one execution instead of one execution per item.
2. Filter early. Put an IF node at the start so spam and duplicates stop before they reach the AI model.
3. Use webhooks, not polling, so workflows only run when there is real work.
4. Pay yearly once you know your plan, to save about 17%.
5. Pick a cheaper model per node. n8n lets each AI node use a different model. Use a small model to sort and a bigger one only to write.
6. Self-host when volume is high and you have someone who can run a server. The Community Edition has no execution limit.

What are the main n8n alternatives?

• Make: also a visual automation tool, but it charges by credits (about one per step), not by full runs. For long workflows with many steps, n8n's per-execution pricing is often cheaper. See the Make cost calculator.
• Zapier Agents: easier for beginners, with a large app library, but usually priced per task or activity, which adds up at high volume. See the Zapier Agents cost calculator.
• Coding on the OpenAI API directly: cheapest to run, as the example above shows, but you need a developer and hosting. See our OpenAI API cost calculator.

Building a phone agent? n8n can handle the follow-up, but you need a voice platform for the call itself. See the Vapi cost calculator.

Who is n8n best for?

What are the key n8n features?

• Visual workflow editor: Connect nodes on a canvas: triggers, apps, logic and AI steps.
• AI Agent node: Lets a model choose which tools to use, with memory and human approval for risky actions.
• Every integration on every plan: Hundreds of app nodes plus HTTP requests to any API.
• Code steps: JavaScript or Python inside a workflow when you need custom logic.
• MCP support: Expose workflows as tools to AI assistants, or call MCP servers from n8n.
• Self-hosting: The free Community Edition runs on your own server with no execution limit.

How do you set up n8n?

The steps below show the general setup. Button names change often, so follow the idea of each step rather than exact labels.

1. Choose cloud or self-hosted. Start a free n8n Cloud trial (no card needed), or install the Community Edition on your own server if you are comfortable running one.
2. Add your AI credentials. Connect your OpenAI, Anthropic or other API key in n8n's credentials so AI nodes can use it.
3. Pick a trigger. Start the workflow from a form, a webhook, a new email, a CRM event or a schedule. Webhooks are best because they only run when there is work.
4. Add an AI Agent node. Write the agent's instructions, pick the model and attach tools such as 'Google Calendar' or 'HTTP Request'.
5. Add memory if it chats. For chat agents, add a memory node so the agent remembers earlier messages in the same conversation.
6. Add a filter before the AI. Use an IF node to stop spam and duplicates before they reach the model and cost money.
7. Test with the editor. Run the workflow with sample data and read each node's output.
8. Activate and monitor. Turn the workflow on, then watch executions and errors. Set up an error workflow that alerts you.

Tips from setup

• Batch items into one execution where possible.
• Use a cheaper model in nodes that only sort or extract data.

How do you calculate your n8n cost?

1. What should your n8n workflow do?. Pick the job closest to yours. We fill in typical usage for that job, so you do not need to know anything about tokens.
2. How many times will it run a day?. Use a normal working day. A rough guess is fine, and we count 22 working days a month.
3. Read your answer. You see one monthly figure, the cost of each n8n task and what makes up the total.

Want to compare more tools at once? Use the main AI agent cost calculator.

How should you budget for n8n?

No-code agent builders charge for runs, credits or activities, and each tool counts them differently. One customer request might be a single run on one platform and a dozen credits on another, so comparing plan prices alone can mislead you. Always convert to the cost of one complete task.

To budget for n8n, build one real workflow and see how much usage a single run consumes. Multiply that by your monthly volume, then add the AI model cost if it is billed separately. Leave room to grow, since crossing a plan limit is what usually makes the bill jump.

n8n pricing FAQ

Get your n8n cost estimate

Now put in your own numbers. The n8n cost calculator at the top of this page gives you a figure in about a minute. Then run a small test with real work and compare your real usage with the estimate. Still deciding? Compare other AI tools.

Source and last checked date

All n8n prices on this page come from the official n8n pricing page, last checked on November 3, 2026. Prices are for monthly billing. Annual billing saves about 17% (Starter $20, Pro $50, Business $800). Business is self-hosted only. Overage fees on Starter and Pro are capped; Enterprise is custom.

More cost calculators

Other no-code agent builders: Make, Zapier Agents, Lindy, Relevance AI.

All tools: OpenAI, Claude, DeepSeek, Gemini, Vapi, Grok, Mistral, GoHighLevel, Retell AI, Bland AI, ElevenLabs, Synthflow, Botpress, Chatbase, Voiceflow. Or go back to the AI agent cost calculator.