n8n
It charges per execution — one run of your entire workflow, whether it has 2 steps or 50. A genuinely different model from Zapier's per-task or Make's per-operation billing, and why n8n runs 3-20x cheaper above 5,000 runs a month.
What is n8n?
n8n, headquartered in Berlin, is an open-source, node-based workflow automation platform genuinely built for technical teams and AI builders who want more control than a drag-and-drop, no-code tool typically offers. The core workspace is a visual editor where you drag triggers (webhooks, schedules, database polling, or a chat trigger for AI agents) onto a canvas, wire them to action nodes across 400-500+ integrations, and drop in transforms like Set, Code, If, Switch, or Merge between them. What genuinely sets it apart technically: its native AI stack includes more than 70 LangChain-based nodes for building agents, managing memory, connecting to vector stores, and making LLM calls — meaning you can wire a genuine ChatGPT or Claude reasoning pipeline directly into the workflow canvas rather than relying on generic HTTP requests the way some competitors do. One detailed 2026 agency report captured the practical impact plainly: switching internal LangChain agents over to n8n cut development time on new agents from roughly 10 hours down to about 1 hour each, though for truly novel agent architectures with custom reasoning chains, hand-written LangChain code still wins.
The genuinely important, structurally distinctive thing worth understanding clearly, especially compared to Zapier and Make, both covered elsewhere in this directory: n8n charges per execution, where one execution is a single run of your entire workflow, no matter how many nodes it contains — a two-step workflow and a fifty-step workflow each burn exactly one execution per run. That's a real, fundamentally different model from Zapier's per-task billing (where every action counts separately) or Make's per-operation billing (where every module action counts), and it's the specific reason n8n runs 3 to 20 times cheaper than Zapier above roughly 5,000 runs a month. The platform's free, open-source Community Edition takes that further: unlimited executions, unlimited workflows, and access to the full integration library, self-hosted with no license fee at all — missing only enterprise features like SSO, LDAP, and advanced permissions. It's honest to note that "free" doesn't mean zero cost in practice: real agency data from two years and 15+ client deployments puts typical self-hosting infrastructure and maintenance costs at $20-100 a month, with one concrete example — a Hetzner CX32 server ($12/month) plus managed PostgreSQL ($15/month) — running 40+ workflows at roughly 200,000 executions a month, a genuinely impressive scale for around $27 in total infrastructure cost.
The execution-based pricing explanation is drawn directly from an independent, dated 2026 guide (TechJack Solutions) specifically written to clarify this model versus competitors. The real agency self-hosting cost data, including the Hetzner example, is drawn from a detailed hands-on breakdown (PxlPeak) based on two years of production deployment. The LangChain AI node details and the 3-20x cost comparison are drawn from a separate independent 14-day hands-on review (StartupOwl).
Key features
Execution-based pricing
One workflow run counts as one execution, regardless of how many steps it contains.
70+ native LangChain AI nodes
Build agents, manage memory, and connect vector stores without leaving the canvas.
Free, unlimited self-hosting
Community Edition runs unlimited workflows and executions with no license fee.
Code nodes
Drop in custom JavaScript or Python logic directly inside any workflow.
400-500+ integrations
Connect Slack, Google Sheets, Airtable, HubSpot, and hundreds more.
10,700+ workflow templates
A large community marketplace covering marketing, sales, AI agents, and DevOps.
Pricing
Community (Self-Hosted)
- Real infrastructure costs typically run $20-100/month
- A $5-7/month VPS via Docker Compose is a common minimal setup
- Missing only enterprise features: SSO, LDAP, advanced permissions
Cloud Starter / Pro
- 2,300-13,700 AI Workflow Builder credits/month, a major 2026 increase
- 14-day free trial, no credit card required
- Annual billing saves 17% versus monthly
Business / Start-up Plan
- SSO/SAML/LDAP, Git version control, 6 shared projects included
- Start-up Plan: 50% off for companies under 20 employees, under €5M funding
- Delivered as a self-hosted license key, renewable annually
AI Workflow Builder credits refresh monthly and don't roll over; your own AI Agent nodes calling OpenAI, Claude, or Gemini bill separately through your model provider's API. Prices reflect n8n's published pricing as of July 2026.
Available models
Integrations & platforms
Pros, cons & best for
Pros
- Execution-based pricing is genuinely, structurally cheaper for complex, multi-step workflows
- Free, unlimited self-hosted Community Edition is a real pricing escape hatch
- Native LangChain AI nodes offer genuinely deeper AI-agent capability than HTTP-based competitors
Cons
- Requires real technical comfort — not the pick for fully non-technical teams
- Self-hosting still carries real infrastructure and maintenance cost, not truly $0
- Setup time (30-60 minutes for self-hosting) is a real trade-off versus zero-setup Cloud tools
Best for
- Technical founders, developers, and AI builders wanting genuine cost control at scale
- Teams building AI agents needing native LangChain-level integration
- Not the pick for fully non-technical teams wanting the absolute simplest setup
Take a look inside
Alternatives
For the simplest possible interface or a visual-first approach instead:
Our verdict
n8n's execution-based pricing model is a genuinely different, structurally superior approach for complex workflows compared to Zapier's task billing or Make's operation billing — a fifty-step workflow costing exactly the same as a two-step one is a real, meaningful advantage once your automations grow past simple triggers, and the free, unlimited self-hosted Community Edition offers a genuine pricing escape hatch neither major competitor can match. Its native LangChain AI nodes reflect a real, deeper commitment to AI-agent building than HTTP-request-based alternatives, with concrete, documented development-time savings to back it up. The honest trade-off: this genuinely requires technical comfort, and self-hosting carries real infrastructure costs and setup time that a fully non-technical team may not want to take on. For developers, technical founders, and AI builders wanting real cost control and deep AI-agent capability, n8n remains one of the strongest, most genuinely well-regarded automation platforms available.
FAQ
How is n8n's pricing different from Zapier's or Make's?
n8n charges per execution — one run of your entire workflow, regardless of how many steps or nodes it contains — while Zapier charges per task (each action) and Make charges per operation (each module action), meaning n8n stays structurally cheaper as workflows grow more complex.
Is n8n's self-hosted Community Edition really free?
The software itself has no license fee and no execution limits, but real infrastructure and maintenance typically cost $20-100/month based on documented agency data — genuinely much cheaper than Cloud plans at scale, but not literally $0.
How much cheaper is n8n than Zapier?
Roughly 3 to 20 times cheaper above about 5,000 executions a month, a direct result of the execution-based pricing model rather than per-task billing.
What makes n8n's AI capabilities different from Make's?
n8n includes 70+ native LangChain-based AI nodes for agents, memory, and vector stores built directly into the canvas, while Make's AI integration relies more on basic HTTP requests — a real, documented advantage for AI-agent development specifically.
Is n8n good for non-technical teams?
Not especially — it requires genuine technical comfort with concepts like nodes, credentials, and debugging execution errors, making it a better fit for developers and technical founders than for fully non-technical business teams.
Is there a discount for startups on n8n?
Yes, companies with fewer than 20 employees and under €5 million in total funding can apply for 50% off the self-hosted Business plan, delivered as a renewable one-year license key.