n8n Review
Reviewed by Carlos Gil · Founder & Editor, The Best Network
Updated 7 min readEditorially reviewed
Quick facts
- Starting price
- Self-host free — Cloud Starter $20/mo (annual) or $24 monthly
- Free plan
- Yes
- Best for
- Engineering and ops teams running serious automation volume
- Platforms
- Self-hosted (Docker/npm), n8n Cloud (web)
BestStacked Score
Weighted average of 5 categories. Methodology below.
The Verdict
n8n is the workflow automation tool we recommend first to engineering and ops teams running serious volume in 2026. The pricing model is the standout: n8n charges per full workflow execution, not per step or per user, so a workflow with 12 steps counts the same as a workflow with 2 steps, and every plan includes unlimited users and active workflows. For teams running multi-step automations at volume, that usually costs less than per-step billing; check Zapier's and Make's current pricing against your own volumes. Add the free self-hosted Community Edition, 500+ integrations, native AI nodes (LangChain, vector stores, OpenAI/Anthropic), and the ability to drop into JavaScript inline when the visual editor isn't enough, and n8n is the most flexible option we review in the category. The honest trade-off: the learning curve is steeper than Zapier's and the visual UX is less polished. For teams past the toy-automation stage, n8n is the one to evaluate first.
Best for
Engineering teams, ops teams, and AI builders running 500+ workflow executions per month.
Skip if
You're a non-technical solopreneur running a handful of simple automations — Zapier is friendlier.
The Background
What Is n8n?
n8n is a workflow automation platform founded in 2019 by Jan Oberhauser in Berlin, Germany. The name stands for 'nodemation' (node-based automation). n8n is fair-code distributed under the Sustainable Use License and the n8n Enterprise License: the Community Edition is free to self-host, while n8n Cloud and the Business and Enterprise licenses are paid. The 2026 product is aimed at engineering and ops teams running serious automation: 500+ pre-built integration nodes (Google Workspace, Slack, Salesforce, HubSpot, Stripe, AWS, the works), native AI nodes for OpenAI/Anthropic/local models with a built-in LangChain integration and vector store support, and the ability to drop into JavaScript or Python anywhere a node doesn't quite fit. The execution-based pricing model (you pay per full workflow run, not per step or per user) favors teams running multi-step workflows at volume: a 12-step workflow counts the same as a 2-step workflow.
The Product
n8n Features: What You Actually Get
500+ integration nodes (and JavaScript anywhere)
Pre-built nodes for Google Workspace, Slack, Notion, Salesforce, HubSpot, Stripe, AWS, OpenAI, Anthropic, Postgres, MongoDB, and 480+ others. When a node doesn't exist or doesn't do exactly what you need, drop into a Code node and write JavaScript or Python inline.
Build a daily workflow that pulls Stripe revenue, joins it with HubSpot deal data via JavaScript, formats it as a Markdown report, and posts it to Slack — all without leaving n8n.
AI agent nodes (LangChain native)
Native LangChain integration with nodes for OpenAI/Anthropic/Mistral/local models, vector stores (Pinecone, Qdrant, Supabase), agent loops, tool calling, and memory. Every Cloud plan includes agents, MCP, and AI nodes, and Starter and Pro can use AI models without bringing your own API keys.
Build a customer support agent that reads incoming Intercom messages, searches a Pinecone vector index of help docs, drafts a response with Claude, and either auto-replies or escalates to a human based on confidence — entirely visually.
Self-host free + Cloud option
n8n is fair-code licensed: the Community Edition is free to self-host on your own infrastructure (Docker, npm, k8s) with almost the complete feature set. n8n Cloud is the managed option for teams that don't want to run infra, and the paid Business plan is self-hosted only.
An ops team self-hosts the Community Edition on its own VPS and pays only for the server, not per execution.
Per-execution pricing
Paid plans charge per full workflow execution, not per step or per user, and every plan includes unlimited users and active workflows. A 12-step workflow that hits Slack, Stripe, HubSpot, and OpenAI counts as one execution. That is the main cost difference from tools that bill each step.
A SaaS company runs a 9-step new-customer onboarding workflow 1,200 times a month, and n8n counts 1,200 executions rather than 10,800 steps.
Webhooks, schedules, queues, and error handling
Trigger workflows from incoming webhooks, on a schedule (cron-style), from app events, or from queue events. Built-in error handling with retry policies, fallback branches, and per-node error workflows. Production-grade orchestration not just point-and-click flows.
Build a webhook receiver for Stripe events that retries 5 times on failure, falls back to a logging workflow on permanent failure, and pages on-call via PagerDuty if errors exceed a threshold.
Projects, roles, SSO, and governance (Business + Enterprise)
Shared projects (1 on Starter, 3 on Pro, 6 on Business, unlimited on Enterprise), admin roles from Pro, SSO and LDAP plus Git version control and multiple environments from Business, and log streaming and external secret stores on Enterprise. An encrypted credential store comes on every plan.
An ops team isolates dev/staging/prod workflows in separate workspaces with RBAC, gives engineers prod read-only access, and audits every credential use for SOC 2 compliance.
Feature recap
- 500+ pre-built integration nodes across SaaS and infra
- Native LangChain + vector store + AI agent nodes
- JavaScript Code node for custom logic anywhere
- Self-host free or use n8n Cloud
- Per-execution pricing (not per task)
- Webhooks, schedules, queues, and error handling
- Shared projects, admin roles, SSO (Business+), log streaming (Enterprise)
- Workflow templates and an active community forum
The Numbers
n8n Pricing: Every Plan Explained
n8n prices by full workflow executions, and every plan includes unlimited users, unlimited active workflows, and every integration. The self-hosted Community Edition is free; Cloud starts at $20/mo billed annually ($24 monthly). Prices are shown in USD to US visitors and in euros in Europe; the US-dollar figures below were re-checked on October 19, 2026.
| Plan | Monthly | Annual | Included Features | Best For |
|---|---|---|---|---|
| Community (self-host) | Free | Free | Free self-hosted edition with almost the complete feature set; you run the infrastructure | Teams comfortable running their own infrastructure |
| Starter (Cloud) | $24 | $20/mo | 2,500 executions/mo, unlimited active workflows and users, 1 shared project, 5 concurrent executions, 1,600 Assistant credits/mo | Solo developers and small teams testing Cloud |
| Pro (Cloud)Best value | $60 | $50/mo | 10,000 executions/mo, unlimited active workflows, 3 shared projects, up to 50 concurrent executions, admin roles, workflow history | Growing teams running production automations |
| Business (self-hosted) | $960 | $800/mo | 40,000 executions/mo, 6 shared projects, SSO and LDAP, multiple environments, Git version control, queue mode; self-hosted only | Companies under 100 employees that self-host |
| Enterprise | Custom | Custom | Custom executions, unlimited shared projects, 200+ concurrent executions, log streaming, external secret stores, dedicated support with SLA; Cloud or self-hosted | Companies running mission-critical automation at scale |
The self-hosted Community Edition is the cheapest route if you can run Docker yourself: you pay for the server, not per execution. Starter at $20/mo billed annually ($24 monthly) suits solo devs and small teams that want managed hosting and 2,500 executions/mo (roughly 80/day). Pro at $50/mo annually ($60 monthly) handles 10,000 executions and is where most working teams land. Business ($800/mo annually, $960 monthly) is self-hosted only and adds SSO, environments, and Git version control. Active workflows are no longer capped on any plan; executions are the meter. Because Zapier and Make bill per step, n8n tends to cost less for long multi-step workflows; compare against their current pricing pages at your real volume.
The Tradeoffs
n8n Pros and Cons
What we love
- Per-execution pricing favors long multi-step workflows
- Self-hosted Community Edition is free
- Native LangChain + AI agent nodes, plus MCP support
- JavaScript anywhere — escape hatch when visual hits limits
- 500+ integrations cover most SaaS and infra needs
- Production-grade error handling and queue management
- Unlimited users and active workflows on every plan
- Fair-code license is clear and enterprise-friendly
What could be better
- Steeper learning curve than Zapier for non-developers
- Visual editor UX is less polished than Make
- Self-host requires Docker/Linux comfort
- Some integrations less deep than Zapier's first-party
- Documentation gaps on newer AI/agent nodes
- Business plan is self-hosted only and costs $800/mo annually
The Audience
Who Should Use n8n?
Best for: engineering and ops teams running serious volume
If you're running thousands of executions a month or building anything multi-step, n8n's per-execution billing usually works out cheaper than per-step billing, and it gives you a JavaScript escape hatch when needed.
Best for: AI builders and prompt engineers
The native LangChain integration, vector store nodes, and visual agent loops make n8n a strong fit for building production AI workflows without writing a full app.
Skip if: you're a non-technical solopreneur
If you're running a handful of simple two-step automations and don't want to think about Docker or JavaScript ever, Zapier is friendlier and the per-task pricing is fine at low volume.
The Methodology
How We Reviewed n8n
Our editorial evaluation considers product information, official documentation, pricing, features and professional experience. It is not a claim of documented hands-on testing.
Our review methodology →What changed
- : Re-checked n8n.io/pricing on October 19, 2026 with both billing toggles: Starter $20/mo annual ($24 monthly, 2,500 executions), Pro $50 ($60, 10,000), Business $800 ($960, 40,000, self-hosted only) and Enterprise (custom) are unchanged in US dollars. Visitors in Europe see the plans priced in euros.
- : Substantive update. Removed the old active-workflow caps (5 on Starter, 15 on Pro): every n8n plan now includes unlimited active workflows and users, metered by executions. Updated plans to Starter ($20/mo annual, $24 monthly, 2,500 executions), Pro ($50/$60, 10,000 executions, up to 50 concurrent), Business ($800/$960, 40,000 executions, self-hosted only), and Enterprise (custom); noted that prices show in USD for US visitors and in euros in Europe. Corrected the Community Edition description, the license (Sustainable Use License), Python support in the Code node, and which governance features sit on which plan. Removed an unverified funding total and customer names, unverified template counts, invented cost-savings and time-savings figures, and unchecked competitor prices.
The Field
n8n vs The Competition
n8n's primary competition is Zapier (the incumbent) and Make (the visual-flow alternative). Our editorial read: n8n leads on flexibility and self-hosting; Zapier on ease of use; Make sits between them.
| Tool | Starting Price | Free plan | Best Feature | BS Score |
|---|---|---|---|---|
| n8n | Self-host free — Cloud Starter $20/mo (annual) or $24 monthly | Yes | 500+ integration nodes (and JavaScript anywhere) | 9.2/10 |
| Zapier | See Zapier | Yes | Non-technical users | 9.4/10 |
| Make (Integromat) | See Make | Yes | Visual flow UX | 8.7/10 |
| Lovable | See Lovable | Yes | Full app generation | 9.1/10 |
Zapier: Zapier is friendlier for non-developers. n8n is more flexible, can be self-hosted, and bills per workflow execution rather than per step.Read our Zapier review →
Make (Integromat): Make has a polished visual editor. It has no self-hosted option, so teams that need to run automation on their own infrastructure should look at n8n.Read our Make (Integromat) review →
Lovable: Different category — Lovable generates full applications from prompts. Compare when you're choosing between 'automate a workflow' (n8n) and 'build a full app with workflows in it' (Lovable).Read our Lovable review →
Use cases
Example setups by buyer type
A B2B SaaS ops team running 12,000 monthly automations
Move multi-step Zapier workflows to self-hosted n8n on a VPS, so the bill is the server rather than per-step usage. The team gains a JavaScript escape hatch and AI nodes; log streaming and external secret stores require an Enterprise license.
An AI product team building a customer support agent
Build a production-grade support agent visually in n8n: incoming Intercom message → embeddings via OpenAI → Pinecone vector search → Claude response generation with tool calling → confidence-based human handoff, with n8n's human approval step for tool calls where needed.
A growth marketer at a 50-person SaaS
Self-serve 14 cross-tool automations on n8n Cloud Pro: lead enrichment from Apollo, deal stage syncing across HubSpot/Salesforce, daily revenue reports to Slack, abandoned-trial reactivation campaigns, all within Pro's 10,000 monthly executions.
The Ecosystem
n8n Integrations
n8n's integration surface is broad (500+ nodes) and includes both SaaS apps and infra primitives, plus the JavaScript Code node when nothing exists.
- →Google Workspace (Sheets, Docs, Calendar, Drive, Gmail)
- →Slack, Discord, Microsoft Teams (messaging)
- →Salesforce, HubSpot, Pipedrive (CRM)
- →Stripe, PayPal, Paddle (payments)
- →OpenAI, Anthropic, Mistral, local LLMs (AI)
- →Pinecone, Qdrant, Supabase, Weaviate (vector stores)
- →Postgres, MySQL, MongoDB, Redis (databases)
- →AWS (S3, Lambda, SES), Google Cloud, Azure (cloud)
- →GitHub, GitLab, Jira, Linear (dev tools)
- →Webhook + HTTP Request (universal)
- →Code node (JavaScript or Python inline)
- →Public REST API, CLI, and workflow templates
The integration most teams underuse: the Code node lets you write inline JavaScript anywhere in a workflow, which means n8n is effectively a backend you can compose visually. Combined with the HTTP Request node, you can integrate literally any API even if no first-party node exists.
Common Questions
n8n FAQ
The Final Word
Should You Use n8n?
n8n in 2026 is the workflow automation platform engineering and ops teams should evaluate first. We score it 9.2/10, held back by a learning curve that's steeper than Zapier's and a visual editor that's less polished than Make's. The per-execution pricing model, with unlimited users and active workflows, favors teams running multi-step workflows at volume. The self-hosted Community Edition is free; n8n Pro at $50/mo billed annually ($60 monthly) is the right tier for most teams that want managed hosting. Pair n8n with Lovable (for full apps), Cursor (for the code that lives in your Code nodes), and the Anthropic API (for agents) and you have a complete modern automation stack.
Yes, if
you're an engineering or ops team running 1,000+ multi-step workflow executions per month
No, if
you're a non-technical solopreneur with five simple two-step automations — Zapier is friendlier
n8n Is Part Of These Stack Packs
About the reviewer
Carlos Gil
Founder & Editor, The Best Network
Carlos Gil is the Founder & Editor of The Best Network, bestselling author of The End of Marketing, and a LinkedIn Top Voice. His digital marketing career began in 2008 and spans corporate roles at LinkedIn and BMC Software, consulting for global brands, creator partnerships, and building his own businesses.
Get the best of BestStacked
The AI tools worth knowing. None of the noise.
- New AI tool and software reviews as they publish
- Head-to-head comparisons and stack breakdowns
- The occasional recommendation worth knowing
No fake urgency. No made-up weekly cadence. Unsubscribe anytime.