A workflow automation platform that combines node-based visual orchestration with inline JavaScript/Python code and native AI Agent nodes. Unlike step-based tools, n8n charges by execution—a single run of an entire workflow, regardless of how many steps it contains. It offers two deployment paths: n8n Cloud (managed, Starter/Pro/Enterprise) and self-hosted (Community Edition free and unlimited, or paid Business/Enterprise editions). The Community Edition is free with unlimited executions, making it popular for teams that want data to stay on their own infrastructure. AI capabilities are deep: built-in LangChain integration, memory components, tool calling, RAG support, and an MCP server that lets AI clients run and edit workflows. Licensing note: n8n uses a fair-code licence (not OSI open source)—internal commercial use is free, but providing n8n as a service to external clients requires a subscription. Pricing shift: the 2025/2026 update moved Cloud and paid self-hosted plans to execution-based billing, with overage costs that scale with volume (roughly $0.015/execution reported).

Pros & Cons:

Pros: The Community Edition is genuinely free and unlimited for self-hosting, which is rare among automation platforms and makes it viable for teams that need data sovereignty; the hybrid node+code architecture lets you start visual and drop into JavaScript or Python when logic gets complex, avoiding the ceiling that pure no-code tools hit; AI Agent nodes are deeply integrated (LangChain, memory, tool calling, RAG), making n8n one of the most capable low-code platforms for building agentic workflows; the built-in MCP server allows AI clients like Claude to run and edit workflows, which few competitors offer; 500+ integrations and an active community reduce the need to build every connector from scratch.

Cons: The execution-based billing shift has caused sticker shock—while the Community Edition remains free, Cloud and Business plans now charge per execution, and high-frequency workflows (polling triggers, chatbots) can burn through allowances quickly; third-party analysis reports overage rates around $0.015/execution, which scales poorly for high-volume use, and failed executions still count; the licence is fair-code, not OSI open source—internal commercial use is free, but reselling n8n as a service requires a subscription, which matters for agencies and SaaS builders; self-hosting requires technical skill (Docker, server maintenance, database configuration), so non-technical users should start with Cloud; SSO, Git version control, and environments are gated behind Business/Enterprise, meaning growing teams face a feature paywall.

Who is it for?:Technical teams that want a self-hosted, fair-code automation platform with unlimited executions and data staying on their own infrastructure; developers building AI Agent workflows who want LangChain-level capability without writing everything from scratch; teams that need 500+ integrations and the ability to drop into code for complex logic; organizations that can absorb the execution-based cost model or stay on the free Community Edition. Not suited for non-technical users seeking a zero-setup SaaS (Cloud exists but is execution-capped), or agencies that want to resell automation services without a commercial licence.