n8n

n8n is a visual workflow automation platform that combines AI agents with business process automation. It offers run history, human approvals, version control and enterprise security, and can be self-hosted with Docker or used as a hosted service.

n8n is a workflow automation platform that puts AI agents and conventional business process automation on the same visual canvas. A workflow is a chain of steps, and any step can be opened to see what it does, so the logic stays readable for people who did not build it. The homepage example follows a refund request: the flow looks up the order, lets a model investigate the claim, asks a manager in Slack for approval, and then either processes the refund or opens a ticket. The product is aimed at technical teams that want the pace of no-code without losing the option to write code. Most assembly happens in the visual builder, and custom code can be added where a ready-made step falls short. Featured use cases include IT onboarding across tools like Slack, Jira and Notion, security threat investigation, alert triage and lead qualification. Customer stories name Vodafone and Huel. Models are interchangeable, and AI actions use structured inputs and outputs, with human approvals, guardrails and evaluations placed around what an agent is allowed to do. Day-to-day operation is built around traceability. Running agents can be watched live, past runs can be opened, and a failed step can be re-run on its own instead of restarting the whole flow. Runs can be replayed or tested with mocked data, and version history records who changed what and when. Larger organisations get git-based control, isolated environments, workflow diffs, SSO via SAML, LDAP, role-based permissions, encrypted secret stores, audit logs and log streaming to a SIEM. Deployment is flexible: use the hosted cloud version, or self-host with Docker or fully on-premises so data stays in your own environment, with the source available on GitHub. Compared with simple connector tools, n8n suits teams comfortable with technical concepts. Compared with code-only agent frameworks, it offers a visual layer that makes automations easier to hand over and audit.