Contains affiliate links. We may earn a commission — it never changes the verdict. Details
The Stack / Automation
n8n
04 AutomationYes for operators who've outgrown per-task pricing — n8n executes whole workflows per credit (or free self-hosted), drops into code when visual logic runs out, and has become the default for serious AI agent workflows. It asks more of you than Zapier; it returns more too.
The job in the stack
The heavy-automation seat: complex, high-volume workflows and AI agent pipelines, self-hosted or cloud.
Pricing
| Plan | Price | Notes |
|---|---|---|
| Community (self-hosted) | $0 | Unlimited executions and workflows, full integration catalog, fair-code license |
| Starter (cloud) | 20€/mo | 20€/mo billed annually, 24€/mo billed monthly — euro figures; US visitors now see the same numbers in dollars ($20/$50/$800 billed annually). 2,500 executions/mo, 5 concurrent, 1 shared project, 1,600 AI Assistant credits/mo |
| Pro (cloud) | 50€/mo | 60€/mo billed monthly. 10,000 executions/mo, 20 concurrent (a 50-concurrent option is offered), 3 shared projects, up to 9,600 AI Assistant credits/mo. Starter and Pro are capped at the execution limit — no overage charges |
| Business (self-hosted) | 667€/mo | 800€/mo billed monthly. Self-hosted only for now — Business on n8n Cloud is waitlist-only. 40,000 executions/mo, SSO/LDAP, 6 shared projects, multiple environments, queue mode, Git version control; AI Assistant listed as a preview. Overage since Oct 26, 2026: 4,000€ ($4,800) per extra 300,000 executions — workflows keep running. Companies under 20 employees can apply for 50% off via the Start-up Plan |
| Enterprise | custom | Unlimited shared projects, 200+ concurrent executions, air-gapped options; AI Assistant preview on request, no standard credit allowance listed |
Pricing verified 2026-10-26. Plans change — the source of truth is n8n's own pricing page.
Models & versions
- AI Agent node v1.28 added structured tool calling, 4 memory backends (in-memory, Redis, Postgres, Motorhead), and a ReAct execution mode; now supports multi-agent orchestration with human-in-the-loop approval gates
The honest read
What works
- + Execution-based pricing: a 50-step workflow costs the same as a 2-step one
- + Self-hosting means data sovereignty and near-zero marginal cost
- + Native AI agent nodes make it a real agent-orchestration layer
- + JavaScript/Python escape hatches when visual builders hit their ceiling
What breaks
- − Steeper learning curve than Zapier — it expects you to think in data
- − Smaller connector catalog; niche SaaS may need HTTP-node handiwork
- − Self-hosting shifts uptime and updates onto you
Use it for
- →High-volume automations where Zapier task math breaks
- →AI agent workflows with branching, memory, and tool calls
- →Teams that want the automation layer inside their own boundary
Skip it if
- ×You need it running today with zero learning curve — start with Zapier
- ×The app you need has no n8n node and you won't touch an HTTP request
Questions operators ask
?Is self-hosted n8n really free?
The software is — you pay in infrastructure and attention. A $5–10/mo VPS runs most solo workloads. If that sentence made sense to you, self-host; if not, cloud Starter.
What changed
- Freshness sweep against n8n.io/pricing (read in the browser and in the page source). Starter 20€ and Pro 50€ per month billed annually hold, as does Business at 667€. US visitors now see dollar pricing: $20 / $50 / $800 billed annually. Real change: Business is self-hosted only, and Business on n8n Cloud is waitlist-only, so we renamed the row. Business overage is listed at 4,000€ / $4,800 per extra 300,000 executions as of today. Starter and Pro stay capped, with no overages. The page also lists Agents (preview), MCP Server Trigger and Client, human approval for tool calls, and AI models without your own API keys on every cloud plan. The monthly-billed figures (24€ / 60€ / 800€) did not render on the toggle in this pass and are carried over.
- Freshness sweep against n8n.io/pricing: AI Assistant credit allowances are lower than we had them. Starter now shows 1,600 credits/mo (we listed 2,300) and Pro shows up to 9,600 (we listed 13,700). Plan prices hold at 20€ / 50€ / 667€ per month billed annually. Pro now offers a 50-concurrent-execution option alongside 20. Business lists the Assistant as a preview with no allowance, and Enterprise lists it on request. Several third-party pricing guides still quote the old credit numbers, so trust the pricing page over them. If you sized a plan around Assistant credits, check it again.
- Freshness sweep against n8n.io/pricing — no changes. Euro pricing, unlimited workflows/users on every cloud tier, and the AI Agent node's memory backends and human-in-the-loop gates all hold as last verified.
- Currency correction: n8n's cloud plans are priced in euros, not dollars — n8n.io/pricing shows 20€/50€/667€ per month billed annually, with no USD toggle. The figures we carried were right; the symbol was wrong. Annual billing saves 17%. Self-hosted Community is still free.
- Freshness sweep: cloud prices held — Starter, Pro, and Business are unchanged, and unlimited workflows and unlimited users stay on every cloud tier, so execution volume is still the only thing that scales your bill. Corrected the Starter row: $20/mo on annual billing, $24/mo monthly. We had been publishing the monthly rate as the annual one.
- Corrected Pro annual price to $50/mo (was listed at ~$60, which is actually the monthly-billed rate). Confirmed Business at $667/mo annual / $800/mo monthly directly on n8n.io/pricing — resolves a standing conflict between sources. AI credit allotments jumped sharply: Starter now 2,300/mo (was 50), Pro up to 13,700/mo (was 150); Business AI Assistant credits are "coming soon," not yet live.
- Verified cloud pricing tiers. AI Agent node upgraded with structured tool calling, four memory backends, and human-in-the-loop approval gates for multi-agent workflows.
- Seeded from January 2026 baseline; cloud pricing pending fresh verification by the update loop.
Give n8n a job
The heavy-automation seat: complex, high-volume workflows and AI agent pipelines, self-hosted or cloud.
Try n8n →Affiliate link — same price for you, keeps the stack maintained.