Is a free process automation tool really free?
That is the first question I asked when I looked at n8n's Community Edition. The claim is straightforward: a self-hosted, source-available workflow automation platform that costs nothing in license fees, no per-task charges, no usage limits. Just infrastructure costs. If you have ever run a server in production, you know that just infrastructure costs does a lot of work in that sentence. I wanted to know what the trade really is, and for whom it actually works.
It’s a trade, not a free lunch
n8n is a node-based workflow builder. You drag nodes onto a canvas, connect them, and each node does one thing: fetch data, transform it, call an API, run JavaScript or Python. It has 400+ native integration nodes — a number that looks small next to Zapier’s 9,000+, but the HTTP node changes that story. The visual editor is clean, and the execution model is simple: a full workflow run counts as one execution, regardless of how many steps it contains. No per-step credit math. You can self-host the Community Edition on your own infrastructure (Docker, Kubernetes, bare metal) or use n8n Cloud (SaaS). The cloud plans also use execution-based pricing. The Community Edition is source-available under the Sustainable Use License — you can self-host, modify, and use it internally, but you cannot offer it as a competing automation service. If you plan to embed this, check with a lawyer.
I have seen teams adopt n8n thinking they would eliminate Zapier costs, only to spend three weeks getting the deployment hardened. That can be the right call if you need data residency or high-volume automation. But call it what it is: an operational investment, not a cost-saving shortcut. If you already run your own infrastructure — if you have a DevOps person, a CI/CD pipeline, and monitoring in place — the marginal cost of adding one more container is near zero. If you are a solo founder or a team without ops experience, the hidden costs can easily exceed a $50/month cloud subscription.
What execution-based pricing actually means at volume
If you choose n8n Cloud, pricing is based on workflow executions — a complete run of a workflow counts as one execution, regardless of how many steps or API calls it makes. That is a real advantage over credit-per-step models if your workflows are step-heavy.
| Plan | Price (billed annually) | Executions per month | Steps per execution |
|---|---|---|---|
| Starter | $20/mo | 2,500 | Unlimited |
| Pro | $50/mo | 10,000 | Unlimited |
For a team running 50,000 workflows per month — say, processing customer orders through a 15-step pipeline — the cost difference with a per-step model like Make's (which charges credits per operation) can be substantial. Execution-based pricing means a complex workflow does not cost more than a simple one. I have seen internal analyses that show n8n Cloud beating Zapier by 30–50% for data-heavy integrations. You can read a full breakdown in our article on process automation tool pricing at scale.
400 integrations sounds small — until you see the HTTP node
Zapier boasts 9,000+ integrations. n8n has 400+. On the surface, that looks like a knockout. But for a developer, the HTTP node changes everything: it lets you call any REST API directly, with full control over headers, authentication, and pagination. Add in the JavaScript and Python code nodes, and you can handle any API workflow without waiting for a pre-built connector.
That said, the smaller library means you will spend more time building integrations yourself. Documentation on newer nodes can be sparse — one review noted that n8n's docs for recently added connectors lag behind the code. If you rely on cutting-edge integrations, expect to dig into the source or community forums.
Where n8n fits hard
When n8n fits, it fits hard. Here is what it genuinely excels at:
- Data control and compliance: SOC 2, RBAC, audit logs, and secret management built in. If your compliance team needs to know exactly where data goes, self-hosted n8n gives you a clean answer.
- Cost predictability at volume: No per-step credits means your monthly bill does not explode when you add a complex new workflow. For teams running tens of thousands of executions, that is a genuine advantage.
- Community and learning: 4,000+ templates, a LinkedIn Learning course by Morten Rand-Hendriksen, and a responsive community forum. The learning curve is real, but the support resources are solid.
- AI capabilities: The AI workflow builder lets you orchestrate LLM calls, build RAG pipelines, and embed AI agents without switching tools. We covered this in our comparison of AI workflow orchestration tools.
User ratings reflect this fit: G2 4.8/5 (131+ reviews), Capterra 4.6/5 (39+ reviews). But those numbers come from a predominantly technical user base. They describe satisfaction among people who already fit the tool's sweet spot. I would not generalize them to non-technical buyers.
Where it falls short: UI performance, docs, and ops drag
The weaknesses are real and worth weighing before you commit:
- UI performance on large workflows: When your canvas has hundreds of nodes, the editor can become sluggish. I have seen reports of noticeable lag at around 150–200 nodes. If your automation involves sprawling logic with many decision branches, test your real workflow before betting the farm.
- Sparse documentation for newer integrations: n8n ships features fast, but the docs team sometimes lags. If you rely on a brand-new node, expect to experiment or read the source.
- Infrastructure maintenance drag: For self-hosted, you own the updates, the backups, the monitoring. A missed update can mean a security gap. This is the cost that most cost–benefit analyses ignore.
- Bring your own API keys for AI: The AI features let you use any LLM provider, but you have to supply and manage the keys yourself. That gives you flexibility and avoids vendor lock-in — but also adds another moving part.
- License restriction: The Sustainable Use License allows internal use and modification, but you cannot offer n8n as a competing automation service. If your use case involves embedding workflow automation into a product you sell, consult a lawyer.
Who should use n8n — and who should skip it
n8n is not a universal recommendation. The division is fairly clean:
| n8n is a strong fit for | n8n is probably not for |
|---|---|
| Teams with infrastructure experience (DevOps, Docker, Kubernetes) | Teams that want a turnkey, zero-ops solution |
| Organizations with strict data residency or compliance requirements | Solopreneurs who cannot afford ops overhead |
| High-volume automation with complex, step-heavy workflows | Simple, low-volume integrations (Zapier free is easier) |
| Developers who prefer custom code over pre-built connectors | Non-technical users who need visual-only configuration |
| Teams that want long-term cost predictability at scale | Organizations that change tools frequently |
For SMBs that need a balance of ease and power, our guide to the best business process automation solutions for small to mid-market teams covers several alternatives that may suit better.
Verdict: the right tool, under the right conditions
I believe n8n's Community Edition delivers the best value for technical teams that already run their own servers and need data control. The execution-based cloud pricing is a genuine advantage for step-heavy workflows. The HTTP node and custom code steps close the integration gap for anyone comfortable with a little code. But I would not recommend n8n to anyone who does not already have operational experience. The 'free' label is accurate only if you count your own time, server costs, and maintenance as zero. If you are that team — if you know what a Docker compose file looks like and you have monitoring in place — n8n is a powerful, cost-predictable platform that will scale with you. If not, the trade is probably not worth it.
The thesis holds, but with a condition: n8n is the right process automation tool for teams that can absorb the infrastructure cost. For everyone else, there are simpler paths.