{"id":"n8n","name":"n8n","homepage":"https://github.com/n8n-io/n8n","repo_url":"https://github.com/n8n-io/n8n","category":"workflow-automation","subcategories":["ai-agents","integration-platform","low-code"],"tags":["workflow-automation","low-code","ai-agents","langchain","self-hosted","400-integrations","typescript","fair-code"],"what_it_does":"n8n is a fair-code workflow automation platform with native AI capabilities, enabling technical teams to build automations using a visual editor combined with custom JavaScript/Python code across 400+ integrations, including LangChain-based AI agent workflows.","use_cases":["Building AI agent workflows with LangChain against custom data and models","Automating business processes across 400+ SaaS integrations without managing custom glue code","Creating data pipelines that combine APIs, databases, and AI processing steps","Self-hosting a Zapier/Make alternative with full code flexibility and data control","Enterprise workflow automation with SSO, advanced permissions, and air-gapped deployment"],"not_for":["Simple two-service integrations where a dedicated connector (Zapier) is faster","Teams without technical staff to manage self-hosted infrastructure","Pure MCP-server use cases — n8n is a workflow platform, not an MCP server itself"],"best_when":"When you need a flexible, self-hostable automation platform that combines visual building with full code access and native AI agent support across hundreds of integrations.","avoid_when":"When you need a lightweight, no-code solution or when your team lacks the capacity to self-host and maintain the platform.","alternatives":["zapier","make","temporal","airflow","prefect"],"af_score":62.2,"security_score":65.0,"reliability_score":null,"package_type":"mcp_server","discovery_source":["github"],"priority":"low","status":"evaluated","version_evaluated":"2.9.4","last_evaluated":"2026-03-01T09:50:05.983336+00:00","performance":{"latency_p50_ms":null,"latency_p99_ms":null,"uptime_sla_percent":null,"rate_limits":null,"data_source":"llm_estimated","measured_on":null}}