fruitpunch

Review · AI & data

n8n logon8n review

A workflow automation platform with AI agent nodes that you can use in n8n's cloud or self-host as the free Community Edition.

By the FruitPunch research desk · Updated November 20, 2026

The short answer

n8n is a workflow automation platform with built-in AI agent nodes that runs either in n8n's cloud or on a server you control, where the Community edition is free. It is a strong pick for technical teams because it bills per full workflow execution with unlimited steps; the cloud Starter plan listed at 20 euros per month billed annually for 2.5K executions when checked in November 2026. Non-technical teams that want the simplest builder are better served elsewhere.

Quality (40%)
9.0
Ease of use (25%)
7.4
Value (35%)
9.0

What it looks like

n8n screenshot: n8n homepage with the visual workflow builder
n8n homepage with the visual workflow builder. Source: official page (n8n.io), November 20, 2026
n8n screenshot: n8n pricing page with cloud and self-hosted plans
n8n pricing page with cloud and self-hosted plans. Source: official page (n8n.io/pricing), November 20, 2026
n8n screenshot: n8n features page covering build, run and govern
n8n features page covering build, run and govern. Source: official page (n8n.io/features), November 20, 2026

Our verdict

n8n suits developers, IT and operations teams that run long, multi-step workflows or AI agents and want predictable costs or the option to self-host. It is a weaker fit for people who have never worked with APIs, JSON or expressions, and for teams that need Git version control, SSO or multiple environments without paying for Business or Enterprise. The deciding fact is the billing unit: n8n's pricing page counts one execution per complete workflow run regardless of how many steps it contains, so a 30-step agent workflow costs the same as a 2-step one.

Overview

n8n is a visual workflow builder for connecting apps, APIs and AI models. Workflows are made of nodes on a canvas, with the inputs and outputs of every step visible, and code steps in JavaScript or Python are available when a no-code node is not enough. n8n's features page says it connects to 500+ apps out of the box, and its HTTP request node covers custom REST and GraphQL calls.

There are two ways to run it. n8n Cloud is the hosted service with Starter, Pro and Enterprise plans. Self-hosting uses the same core product: without a license key it runs as the free Community edition, a free registration adds folders, debug in editor and custom execution data, and Business or Enterprise license keys enable team features. n8n's docs describe a one-line installer that needs Docker on Linux or macOS, plus Docker Compose, npm and cloud provider options.

The current plan line-up centers on AI agents. Every cloud plan includes the AI Agent node and AI steps, an MCP Server Trigger, an MCP client, instance-level MCP, human approval for tool calls and hosted or embedded chat. Cloud plans also include monthly Assistant credits for n8n's AI building assistant, which is still labeled preview: 1,600 on Starter and 4,000 or 9,600 on Pro depending on plan size. The pricing page also notes a Business overage fee change on 26 October 2026.

Compared with the other agent platforms on this site, n8n trades ease of use for control. n8n's pricing page contrasts its model with tools that charge per step or per user, which matters most for long workflows. Zapier Agents and Make are simpler starting points for non-technical teams, Dify is the closer match for teams focused on LLM apps and retrieval, and Microsoft Copilot Studio fits organizations already standardized on Microsoft 365. n8n stands out on self-hosting and on per-run billing.

The limits are real. The cloud Starter plan allows 5 concurrent executions, a 5-minute maximum execution time and 1 shared project. Git version control, environments, SSO and LDAP are listed under Business and Enterprise, not the cloud Starter or Pro plans. The self-hosted Community edition has no projects, no workflow or credential sharing and no SSO. The code is published under n8n's Sustainable Use License, a fair-code license that allows use for internal business purposes but needs a separate commercial agreement for some uses, such as offering n8n to others as a paid service.

From the location where prices were checked, the n8n pricing page listed plans in euros only, so the plan prices below are shown in euros. US visitors may see different figures in US dollars.

How it scored

n8n ranks second because it pairs a capable agent toolkit with a pricing model that stays fair as volume grows. Every cloud plan includes the AI Agent node, AI steps, an MCP server trigger, an MCP client, hosted chat and human approval for tool calls, and n8n bills per workflow execution regardless of how many steps it has. All plans include unlimited users, unlimited workflows and every integration. The Starter cloud plan includes 2.5K executions with 5 concurrent runs; Pro raises that to 10K executions and up to 50 concurrent runs. Teams that can run their own server can use the Community Edition from GitHub instead. The trade-off is setup: building agents here feels closer to programming than Zapier does, and features such as Git version control, environments and SSO sit on Business and Enterprise. From our location the pricing page listed euro prices only, so we marked the USD price as pending.

Who it is for

technical teams that want self-hosting and per-run pricing

Good fit if

  • Developers and IT teams who want to self-host automation on their own infrastructure at no license cost with the Community edition
  • Operations teams running long multi-step workflows, where per-step billing on other tools would add up quickly
  • Teams building AI agents that need MCP support, tool calling with human approval and a choice of models
  • Agencies and growing companies that need unlimited users and workflows on every plan
  • Organizations with data residency or compliance needs that can run n8n on-premises or in a private cloud

Skip it if

  • Non-technical users who want plain-language automation with minimal setup: Zapier Agents is easier to start with
  • Small teams that need Git version control, environments or SSO on a tight budget: those features start at Business, so Make may cost less for simple scenarios
  • Teams focused only on chat apps and retrieval over documents: Dify is built around LLM apps and knowledge bases
  • Companies that want to resell automation as a paid hosted service without a commercial agreement: the Sustainable Use License does not allow it

Specifications

From n8n's own pages, each linked

Billing unit
Workflow executions, unlimited steps per execution
source
Users and workflows
Unlimited on all plans
source
Cloud Starter allowance
2.5K workflow executions per month, 5 concurrent executions
source
Cloud Pro allowance
10K workflow executions per month, 20 or 50 concurrent executions
source
Max execution duration
5 minutes on Starter, 40 minutes on Pro and Enterprise
source
Integrations
500+ apps, plus custom HTTP and GraphQL requests
source
Code steps
JavaScript and Python on all plans
source
AI agent tooling
AI Agent node, MCP Server Trigger, MCP client, human approval for tool calls
source
Saved executions
2.5k on Starter, 25k on Pro, 50k on Enterprise
source
Workflow history
1 day on Starter, 5 days on Pro, 365+ days on Enterprise
source
Self-hosting
Free Community edition; Docker, Docker Compose, npm or one-line setup
source
License
Sustainable Use License and n8n Enterprise License (fair-code)
source
Compliance
SOC 2 Type II
source
Free trial
Cloud trial with no credit card: 1,000 executions, 180-second timeout
source

Price

See vendor pricing (USD check pending).

US price check pending. See the vendor pricing page.

Plans and prices

Quoted from the seller, with the date we checked

n8n plans and prices with sources and check dates
PlanPriceWhat is includedChecked
Community edition (self-hosted)Freeno license fee; you pay for your own serverAlmost the complete n8n feature set, per n8n's docs; No projects, sharing, SSO, environments or Git version control; Free registration adds folders, debug in editor and custom execution dataChecked Nov 20, 2026
Starter (cloud)20 euros per month (USD price not shown from our location)billed annually; monthly billing also offered, annual saves 17%2.5K workflow executions with unlimited steps; 1 shared project and 5 concurrent executions; Agents, MCP and AI nodes; 1,600 Assistant credits per month; Forum supportChecked Nov 20, 2026
Pro (cloud)50 euros per month (USD price not shown from our location)billed annually; monthly billing also offered, annual saves 17%10K workflow executions with unlimited steps; 3 shared projects and up to 50 concurrent executions; 7 days of insights and workflow history; Up to 9,600 Assistant credits per month; Admin roles and metric-based evaluationsChecked Nov 20, 2026
Business (self-hosted)667 euros per month (USD price not shown from our location)billed annually40K workflow executions with unlimited steps; 6 shared projects, SSO and LDAP; Multiple environments, queue mode and version control using Git; 30 days of insights; Startup plan: 50% off Business for eligible companies under 20 employeesChecked Nov 20, 2026
Enterprise (cloud or self-hosted)price not listed, contact salescustom contract, pay by invoiceCustom workflow executions and 200+ concurrent executions; Unlimited shared projects and 365 days of insights; External secret store, log streaming and audit logging; Dedicated support with SLAChecked Nov 20, 2026

Key features

  1. Per-execution billing

    n8n counts one execution per complete workflow run, regardless of the number of steps or the amount of data processed. Starter and Pro are capped with no overage fees, so workflows that hit the limit do not generate surprise charges.

  2. AI agents and MCP

    The AI Agent node, an MCP Server Trigger, an MCP client and instance-level MCP are included on every cloud plan. Tool calls can require human approval before they run, and agents can be published as hosted or embedded chat.

  3. Self-hosting

    The same product runs on your own servers via Docker, Docker Compose, npm or a one-line installer. Without a license key it runs as the free Community edition, and n8n documents a self-hosted AI starter kit with Ollama, Qdrant and PostgreSQL for running AI with no external data transfer.

  4. Code when needed

    Code steps accept JavaScript or Python, the HTTP node handles custom REST and GraphQL calls, cURL commands can be imported, and n8n itself can be controlled through an API and CLI. Bash scripts and custom nodes are available on self-hosted installs.

  5. Debugging and replay

    Executions are logged with inputs and outputs for every step, failed runs can trigger error workflows, and past executions can be loaded back into the editor. Pro adds search of executions by the data they contain.

  6. Governance on higher tiers

    Business and Enterprise add SSO and LDAP, Git-based version control, separate environments and queue mode for scaling. Enterprise adds external secret stores such as AWS Secrets Manager, Azure Key Vault and HashiCorp Vault, audit logging and log streaming.

Strengths and weaknesses

What we like

  • Execution-based billing with unlimited steps: a long agent workflow counts as 1 execution, while n8n notes that other tools charge per step or user
  • Free self-hosted Community edition with almost the full feature set and no execution limit set by a plan
  • Unlimited users and unlimited workflows on every plan, including the 20 euro Starter plan
  • Agent tooling on all cloud plans: AI Agent node, MCP server and client, and human approval for tool calls
  • Starter and Pro have capped usage with no overage fees, according to the pricing page
  • JavaScript and Python code steps on every plan for logic no-code nodes cannot handle
  • Startup plan offers 50% off Business for eligible companies under 20 employees

What could be better

  • Starter limits are tight: 5 concurrent executions, a 5-minute execution cap and 2.5K executions per month
  • Git version control, environments and SSO are not on cloud Starter or Pro; they start at Business, listed at 667 euros per month billed annually
  • The free Community edition has no workflow or credential sharing, so only the owner and the creator can access each workflow
  • The Sustainable Use License is not an OSI open source license and requires a commercial agreement for some uses, such as offering n8n as a paid service
  • Building with nodes, JSON data and expressions has a steeper learning curve than plain-language builders
  • The pricing page showed only euro prices from our location, so USD buyers should confirm their price at checkout

Key facts

  • Free option: Self-hosted Community Edition
  • Billing unit: Workflow executions, unlimited steps each
  • Starter allowance: 2.5K workflow executions
  • Users: Unlimited on all plans
  • Startup discount: 50% off Business for companies under 20 employees, if eligible

Where it ranks

n8n is No. 2 in our Best AI Agent Platforms verdict.

Alternatives to consider

Zapier Agents

Easier for non-technical teams that want a simpler builder; n8n's own pricing page contrasts its per-run billing with tools that charge per step or user.

Make

A visual scenario builder that many teams find gentler than n8n for simple automations; compare its usage-based pricing against n8n's per-run model.

Dify

Better suited to teams building chat apps and retrieval over their own documents, and also offers a self-hosted option.

Microsoft Copilot Studio

The natural choice for organizations standardized on Microsoft 365 that want agents inside Teams and Microsoft's admin controls.

Questions readers ask

Is n8n free?
Yes, the self-hosted Community edition is free with no license fee, though you pay for your own server. n8n Cloud has no free plan, but it offers a free trial with no credit card that includes 1,000 executions.
How much does n8n cost?
When checked in November 2026 from Europe, n8n Cloud Starter was listed at 20 euros per month and Pro at 50 euros per month, both billed annually. Self-hosted Business was 667 euros per month billed annually, and Enterprise pricing is by quote. US prices may differ.
Is n8n worth it?
n8n is worth it for technical teams running multi-step workflows, because it bills per full workflow run with unlimited steps and can be self-hosted for free. Teams without anyone comfortable with APIs or JSON may get more out of a simpler tool such as Zapier.
n8n vs Zapier: which is better?
n8n is the better fit for complex, high-volume or self-hosted workflows because it counts whole workflow runs instead of individual steps and offers a free self-hosted edition. Zapier is the easier choice for non-technical users who want quick app-to-app automations.
Is n8n open source?
No, not in the OSI sense: n8n uses the Sustainable Use License, a fair-code license, plus the n8n Enterprise License for some files. The source code is public on GitHub and free to use for internal business, personal or non-commercial purposes.
What is an execution in n8n?
An execution is one complete run of a workflow, no matter how many steps it has or how much data it processes. The cloud Starter plan includes 2.5K executions per month and Pro includes 10K.
Can n8n build AI agents?
Yes, every cloud plan includes the AI Agent node, AI steps, an MCP Server Trigger and an MCP client. Tool calls can require human approval, and agents can be published as hosted or embedded chat.
What do you need to self-host n8n?
n8n's one-line setup needs Docker on Linux or macOS, or WSL on Windows. Docker Compose is the documented route for production deployments with databases and other services.

Sources we checked

  1. n8n Plans and Pricing, checked November 20, 2026
  2. n8n Features, checked November 20, 2026
  3. Host n8n, n8n Docs, checked November 20, 2026
  4. Compare editions, n8n Docs, checked November 20, 2026
  5. Community license, n8n Docs, checked November 20, 2026

Deep review last updated November 20, 2026. We recheck prices and specs against these pages; send corrections to the address below.

Alternatives we scored

Visit n8n.io · Questions or corrections: editorial@fruitpunch.ai