n8n Pricing: Free vs Cloud vs Self-Hosted

n8n is unusual in the automation space because it gives you a genuine choice of deployment model. Most SaaS tools give you one option: pay their price, on their infrastructure. n8n gives you three: self-host it free, use their managed cloud on a budget, or scale into enterprise. Each is right for a different operator, and picking the wrong one costs you either money or time.

This breakdown covers what each tier actually costs, what you're giving up or gaining, and which one makes sense for specific use cases. No filler. Just the numbers and the trade-offs.

What n8n Actually Is (Briefly)

n8n is an open-source workflow automation platform. Think Zapier or Make, but with a visual node editor, full code access when you need it, and pricing that doesn't penalize you for running thousands of automations. It integrates with 400+ services and supports LLM nodes, so you can build agentic pipelines where data flows through Claude or GPT-4o as part of a larger workflow. That's where it gets genuinely interesting next to the older automation platforms.

Option 1: Self-Hosted (Free)

n8n is open source under the Sustainable Use License. That means you can run it on your own infrastructure at zero software cost. What you pay instead is setup time and ongoing ops overhead.

What you need: A server. A $5–$10/month VPS (Hetzner, DigitalOcean, Railway) is sufficient for most solo and small-team use cases. Docker is the recommended install method and takes about 20 minutes if you've done it before. Add a reverse proxy (Caddy or Nginx) for HTTPS, and you're done.

What you get: Full n8n, unlimited workflows, unlimited executions, all integrations. No execution limits. No workflow caps. No usage-based billing. Everything that exists in the cloud version is available self-hosted, including the AI agent nodes.

What you lose: Managed updates, automatic backups, uptime SLAs, and support. If your server goes down, your automations stop. You're on the hook for patching, database management, and monitoring.

Right for: Developers, technical founders, and anyone comfortable with a VPS. If you run more than $20/month in automations on Zapier and have basic Linux confidence, self-hosting pays for itself immediately.

Total cost: $0 (n8n license) + $5–$10/month (VPS) = $5–$10/month

Option 2: n8n Cloud

n8n Cloud is the fully managed version: no server, no setup, automatic updates, and n8n owns the infrastructure. Three paid tiers:

The Starter plan at $20/month covers a surprising amount for solo use. 2,500 executions handles most content pipelines, CRM sync tasks, and daily notification workflows. Where it gets tight is high-frequency automations: poll an API every 5 minutes or process webhooks from a busy system, and you'll hit the limit faster than expected.

The Pro plan at $50/month unlocks 10,000 executions and 3 users, which makes it a real team tool. The 15 active workflow cap is the more likely constraint for heavy users, who often hit it before the execution ceiling.

Right for: Non-technical operators who want automation without server management. Agencies and small teams on Pro. Anyone where downtime has a real cost and they'd rather pay $20/month than manage a VPS.

Execution Count: What Actually Burns Your Quota

One thing n8n's pricing page doesn't make obvious: an "execution" is one complete workflow run, not one action. A workflow with 15 nodes that runs once = 1 execution. This is dramatically better than Zapier's per-task pricing, where the same 15-step workflow would cost 15 tasks.

Where quota disappears fast: error retries, high-frequency triggers (polling every minute), and webhook-heavy workflows that receive bursts of events. If you're building an agentic pipeline where Claude processes inbound messages and triggers follow-up workflows, count executions carefully, because each sub-workflow call increments the counter.

n8n vs Zapier vs Make: Real Cost at Scale

Make wins on raw entry price and holds it at scale: the $9/month Core plan already covers 10,000 operations. Zapier loses badly on scale, climbing to roughly $600/month for the same 10,000 tasks once overage pricing kicks in. n8n wins once you factor in self-hosting: at $10/month on a VPS, you run unlimited executions at a price that doesn't move as your workflow count grows. That flat curve is the whole leverage argument.

The decision matrix: if you're non-technical and want the lowest monthly spend with decent power, Make at $9/month. If you're technical and want maximum flexibility at minimum cost, n8n self-hosted. If you want managed n8n without ops work, n8n Cloud at $20/month is fair value.

The AI Agent Nodes: Where n8n Gets Interesting

n8n added native AI agent nodes that let you run LLM-powered reasoning inside your workflow. You can chain a Claude API call to extract structured data from a document, pass it through a conditional, and route it to different endpoints based on the model's output, all without writing backend code.

This matters for solopreneurs and small teams building content workflows, customer support triage systems, or lead scoring pipelines. The AI nodes work on both self-hosted and cloud versions. You bring your own API key for the LLM (Claude, OpenAI, etc.), so the AI cost is separate from n8n's pricing. Budget $5–$30/month in API usage depending on workflow volume. Read our full n8n review for a look at how the AI agent nodes perform in real agentic workflows.

Which Option Should You Choose?

Self-hosted: You're technical, want zero software cost, and have time to manage a VPS. Best value long-term.

n8n Cloud Starter ($20/month): You want managed infrastructure, you're a solo operator, and 2,500 executions/month covers your actual usage. Check if there's a current n8n discount before you commit to the annual plan.

n8n Cloud Pro ($50/month): You're a small team, need collaboration, or your workflows are high enough volume that the Starter limits bite you regularly.

Enterprise: you need SSO, audit logs, and a formal SLA. Get a quote from n8n directly, since pricing is negotiated.

FAQ