N8N

n8n reached $40 million ARR in July 2025 with a 67-person team — an impressive revenue-per-employee ratio. The company raised $180 million in Series C funding, achieving a $2.5 billion valuation. n8n has 400+ native integrations (growing to 500+ with community additions), more than 6,740 community AI automation workflows in the workflow library, and a growing community of developers and enterprises choosing it as a Zapier and Make alternative.

n8n Community Edition is free, open-source, and self-hosted — no execution limits, no workflow limits, your infrastructure. n8n Cloud is the managed SaaS version ($20/month Starter, $50/month Pro) — Grafana-managed infrastructure, automatic updates, SSL, and team collaboration features. n8n Enterprise (pricing negotiated) adds SSO/SAML, role-based permissions, audit logs, and dedicated support. For most technical teams, Community Edition self-hosted on a $20/month VPS is the best value. n8n Cloud suits teams wanting managed deployment without infrastructure overhead.

n8n significantly outperforms Zapier for AI agent automation. n8n's AI Agent node provides LLM + tool calling: the AI decides which of your defined tools (HTTP APIs, database queries, sub-workflows) to call, in what order, and iterates until the task is complete — true autonomous agent behavior with persistent memory. Zapier's AI features primarily offer AI-generated workflow descriptions and basic OpenAI integration in a linear step model — not agent-style tool calling loops. For teams building AI-powered automation, n8n is the clear choice.

n8n uses the Sustainable Use License (fair-code) — free for internal use and non-competing commercial use. Key restriction: you cannot offer n8n as a managed service or embed it in a product that competes with n8n itself. For 99% of use cases (internal business automation, client automation delivery), n8n is effectively free and open-source. The license is designed to prevent cloud providers from reselling n8n as a service without contributing — typical enterprise and agency use is unrestricted.

Standard n8n self-hosted deployment: Docker Compose with n8n service + PostgreSQL database (required for production; SQLite is dev-only). SSL via Nginx reverse proxy with Let's Encrypt or Cloudflare. Environment variables for DB credentials, encryption keys, and webhook URL. For higher availability, n8n on Kubernetes with a 2-replica Deployment and a managed PostgreSQL (RDS, Cloud SQL). We provide Docker Compose configurations, Kubernetes Helm values, and infrastructure-as-code templates for standard deployment patterns.

n8n's AI Agent node creates autonomous LLM agents with tool calling. Configuration: (1) choose model (OpenAI GPT-4o, Claude Sonnet, Gemini Flash); (2) define tools (n8n nodes become tool functions the AI can call — HTTP requests, database queries, workflow triggers); (3) set system prompt and agent memory type (buffer memory for conversation history, external DB for cross-workflow persistence). The agent receives a task, decides which tools to call, executes them, incorporates results into its reasoning, and continues until complete — multi-step autonomous execution.

n8n manages OAuth 2.0, API keys, and bearer tokens for 500+ services through its Credentials system — configured once in n8n settings, referenced by multiple workflows. OAuth tokens auto-refresh without manual intervention. Custom API authentication via HTTP Request node header configuration or pre-authentication Code node logic. Credentials are stored encrypted in n8n's database with AES-256 encryption. We configure credentials as part of every workflow deployment, never exposing secrets in workflow node configurations.

Migration process: (1) audit all existing Zapier workflows — complexity, execution volumes, and cost; (2) identify which workflows are direct node-to-node equivalents vs which require Code node logic; (3) rebuild workflows in n8n using equivalent native nodes where available, Code nodes for custom logic; (4) test with sample data in n8n's execution environment; (5) run parallel for 2 weeks verifying outputs match; (6) cutover and monitor. Most clients see 70-90% cost reduction after migration. Simple linear Zaps rebuild in 30 minutes; complex multi-step workflows with conditionals take 2-4 hours each.

Yes — n8n's HTTP Request node connects to any REST API, GraphQL endpoint, or webhook without a native integration. You configure: URL, method, headers (including authentication), request body, and response parsing. For APIs requiring OAuth not in native nodes: use the HTTP Request node with manual OAuth token management or a Code node for custom auth flows. This means if a service has an API, n8n can integrate with it — the 500+ native nodes just add OAuth management and UI-friendly field mapping on top of what HTTP Request provides.

We provide n8n managed support: workflow monitoring and error alert response, updates when n8n releases new versions (upgrade planning and testing), integration updates when third-party APIs change, new workflow development as automation needs evolve, and infrastructure maintenance for self-hosted deployments. We also offer n8n training sessions for in-house teams building their own workflows, and quarterly automation audits identifying manual processes that n8n can automate.