n8n
n8n in 2026 adds native MCP client/server support and OpenTelemetry tracing. Cloud plans run from a free self-hosted tier to Business at 667 euros/mo.
About n8n
What is n8n (2026 update)
n8n is a workflow automation platform that lets technical teams build multi-step automations by connecting apps, APIs, databases, and AI models on a visual canvas, with the option to drop into JavaScript or Python code nodes for custom logic. It is positioned between simple no-code tools and fully custom backend integration work, and it can be run as a managed cloud service or self-hosted.
What is new in 2026
n8n added native support for the Model Context Protocol, so a workflow can act as both an MCP client and an MCP server, which lets n8n plug directly into the growing ecosystem of MCP-compatible AI tools instead of relying only on custom HTTP nodes. The AI Agent node now validates every tool call against a JSON schema and forces the model to retry if it sends malformed data, which cuts down on silent agent failures. On the infrastructure side, n8n now emits OpenTelemetry traces for workflow executions so runs show up in an existing observability backend without extra exporters, and Task Runners are enabled by default for Code nodes so a runaway script executes in an isolated process instead of crashing the main instance. Error handling also got a rework in 2026, with per-node retry and exponential backoff, a dedicated error branch that can route failed items to a separate workflow, and a new Error Trigger node for failure notifications.
Key features
- Visual workflow builder with hundreds of app integrations plus generic HTTP, code, and database nodes
- Pay-per-execution pricing model, where a full workflow run counts once regardless of how many nodes it contains
- Native MCP client and server support for connecting to and exposing AI tools
- AI Agent node with JSON schema validation on tool calls
- Per-node retry with exponential backoff, error branches, and an Error Trigger node
- OpenTelemetry tracing and an insights API endpoint for execution success, failure, and time-saved metrics
Pricing in 2026
n8n Cloud is billed in euros and priced annually by default. All plans include unlimited users, unlimited workflows, and access to all integrations, with the difference coming from execution volume, concurrency, and enterprise controls. A self-hosted Community Edition is free and open source with no execution limits.
| Plan | Price | What you get |
|---|---|---|
| Community (self-hosted) | Free | Open-source self-hosted install, no execution limits, core features, no official support |
| Starter | Approx. 20 euros/month (billed annually) | 2,500 executions/month, 1 shared project, 5 concurrent executions, forum support |
| Pro | Approx. 50 euros/month (billed annually) | 10,000 executions/month, 3 shared projects, 20 concurrent executions, admin roles, global variables |
| Business | Approx. 667 euros/month (billed annually) | 40,000 executions/month, 6 shared projects, self-hosting option, SSO/SAML/LDAP, Git version control, 30-day insights |
| Enterprise | Custom | Custom execution volume, unlimited projects, 200+ concurrent executions, 365-day insights, dedicated SLA support, log streaming |
Cloud pricing is quoted in euros on n8n’s own pricing page, so US dollar billing amounts will vary with the exchange rate, check the official pricing page for the current numbers in your billing currency. Companies with under 20 employees can apply for a startup discount of roughly 50% off the Business plan. A free trial with Pro-level features and a 1,000 execution cap is available on Cloud with no credit card required.
Who should use it
n8n suits technical teams, agencies, and developers who want more control than a pure no-code tool like Zapier offers but do not want to build integration logic from scratch. The code nodes, self-hosting option, and now native MCP support make it a strong fit for teams building AI agent workflows that need to call external tools reliably, and the per-execution pricing tends to favor workflows with many steps run relatively few times, compared to per-step billing.
Limitations and alternatives
Self-hosting n8n gives full control and no execution caps, but it puts hosting, updates, and scaling on the team running it, which is a real ongoing cost even though the software itself is free. Cloud pricing is quoted in euros and structured around execution volume and concurrency, so teams running many small workflows in parallel should model their expected execution count before picking a tier, since jumping from Pro to Business is a large price step. If n8n does not fit, alternatives include Zapier for simpler no-code automations with a larger prebuilt app catalog, Make (formerly Integromat) for a similarly visual but more usage-metered pricing model, and directly building on a serverless platform for teams that want full custom control without any workflow-tool layer at all.
Frequently asked questions
Is n8n free to use in 2026?
Yes, the Community Edition is free and open source and can be self-hosted with no execution limits, though it lacks official support and some enterprise controls.
Does n8n support the Model Context Protocol in 2026?
Yes, n8n added native support for MCP in 2026, letting a workflow act as either an MCP client or an MCP server.
How is n8n Cloud priced?
n8n Cloud plans are billed annually in euros and scale by monthly execution volume, from Starter through Pro and Business up to custom Enterprise pricing. Check the official pricing page for exact current rates.
How does n8n's pricing model differ from step-based tools?
n8n counts a full workflow run as one execution regardless of how many nodes or steps it contains, rather than billing per individual step.
Information last verified: September 2026.
Related Tools
