n8n is a workflow automation platform from Germany. Using both a visual UI — where you connect building blocks called nodes on a canvas — and code written in JavaScript or Python, you can build complex business workflows and AI agents. It connects to external services through 500+ pre-built nodes, supports multiple AI models and RAG systems, and offers MCP (Model Context Protocol) integration. Its source code is published under a fair-code license (Sustainable Use License), and in addition to the cloud version you can self-host it via Docker or on-premises — a major difference from fully SaaS-based tools like Zapier and Make.
Key Features
- Hybrid of visual UI and code: You can build workflows entirely by dragging and connecting nodes, but you can also drop in JavaScript or Python code exactly where you need it. It combines no-code ease with the freedom of code
- 500+ pre-built nodes: Integration nodes are available for major services such as Slack, Google Sheets, Notion, GitHub, and various databases. With the HTTP Request node, you can also connect to any API that lacks a dedicated node
- Built for AI agents: You can switch between multiple AI models from OpenAI, Anthropic, Google, and others, and build RAG systems combined with vector databases or AI agents that autonomously call tools — all as workflows
- MCP (Model Context Protocol) support: You can expose n8n workflows as MCP servers, or call external MCP servers as tools for your AI agents
- Self-hostable: The source code is published under a fair-code license (Sustainable Use License), and you can run it for free on your own servers via Docker or on-premises. This makes it a good fit for projects where data cannot leave your own environment
- Execution history and debugging: For each workflow run, you can inspect the input and output data of every node, making it easy to pinpoint errors and re-run workflows
Pricing
| Plan | Monthly Price (billed annually) | Executions/month | Key Features |
|---|---|---|---|
| Community (self-hosted) | Free | Unlimited (depends on your environment) | Run the GitHub-published source on your own server |
| Starter | €20 | 2,500 | 1 shared project, 5 concurrent executions, unlimited users |
| Pro | €50 | 10,000 | 3 shared projects, 20 concurrent executions, admin roles, workflow history |
| Business | €667 | 40,000 | Self-hosted, SSO/SAML/LDAP, separate environments, Git version control |
| Enterprise | Contact sales | Custom | 200+ concurrent executions, external secret store, support with SLA |
Pricing is as of August 2026. For the latest pricing, please check the official website.
Pros & Cons
✅ Pros
- Self-hosting lets you run workflows without worrying about per-execution billing, and keeps sensitive data inside your own environment
- You can start with no-code and gradually write code only for the complex parts
- AI features such as agents, RAG, and MCP are built into the platform itself, so you can try them without extra tools
- Workflows can be exported as JSON, making Git-based version control and community template sharing easy
⚠️ Cons
- Compared to Zapier or Make, the learning curve is steeper, and it can be challenging for non-engineers to master on their own
- Self-hosted operation requires you to handle server management, updates, and backups yourself
- The fair-code license restricts certain uses such as reselling it as a commercial SaaS, and it is not OSI-approved open source
- While the node catalog is large, dedicated nodes for Japan-specific services (accounting, groupware, etc.) are scarce, often requiring direct API connections
Comparison with Similar Services
| Comparison | n8n | Zapier | Make | Dify |
|---|---|---|---|---|
| Main use | Workflow automation + AI agents | Automated connections between services | Visual business automation | Building AI apps and agents |
| Target users | Engineers to technically-minded business users | Mainly non-engineers | Non-engineers to intermediate users | AI app developers |
| Self-hosting | Yes (free Community edition) | No | No | Yes (OSS edition available) |
| Code support | JavaScript/Python | Limited | Limited | Yes |
| Billing unit | Workflow executions | Tasks | Operations | Messages/resources |
Who Is It For
- Engineers who want to start with routine task automation and eventually handle AI agents and RAG on a single platform
- Companies that cannot entrust customer or internal data to external SaaS tools and want a self-hosted automation platform
- Teams running high-volume automations for whom Zapier’s per-task billing no longer makes economic sense
- Technically-minded business users frustrated by the limits of no-code tools who want to extend workflows with code only where needed
Summary
n8n is a workflow automation platform that combines the approachability of a visual UI with the freedom of code. With AI agents, RAG, and MCP support built into the platform, it is increasingly straddling the boundary between automation tools and AI infrastructure. The self-hosting option makes it a particularly strong candidate for teams that prioritize data sovereignty and cost control. On the other hand, getting the most out of it requires some technical background, so if you want a fully no-code experience, we recommend comparing it with Zapier or Make.