n8n

n8n is a workflow automation and AI agent builder for technical teams, run in n8n's cloud or self-hosted, and priced by workflow executions rather than seats.

n8n connects a company's apps, APIs and AI models in workflows, so AI agents run inside the processes that already move its data. n8n GmbH, founded by Jan Oberhauser in 2019, develops it from its Berlin headquarters. Each workflow is a chain of nodes on one canvas, and an agent is one node in that chain. The same software runs in n8n's cloud or on a company's own servers through the self-hosted Community edition. The AI Agent node takes a chat model from the provider the team chooses, a memory for the conversation and a set of tools. Tools can be another workflow through the Call n8n Workflow Tool, an HTTP request, or a function written in the Custom Code Tool. The AI Agent Tool lets a root agent hand work to other agents, and vector store nodes give each agent documents to retrieve from. The Structured Output Parser holds replies to a JSON Schema, while the Guardrails node checks text going into and out of a model. The MCP Client node uses tools over the Model Context Protocol, and the MCP Server Trigger offers a workflow to other AI clients. n8n suits technical teams that build their own automations, from solo builders with workflows in production to organisations with strict compliance and governance needs. Companies that must keep data on their own infrastructure can self-host it, and the Community edition runs without a subscription. Every plan includes unlimited users and workflows, and the bill follows the number of workflow executions, however many steps each one has. n8n hosts the Starter and Pro plans, the Business plan is for self-hosted companies, and Enterprise is arranged with sales in either form. A support team can put the Chat Trigger's embedded chat on its website and ground the agent in documents held in Qdrant. An operations team can start a workflow from a Webhook node, have an agent draft a reply and wait for a Slack approval. Developers return an agent's answer through the Respond to Webhook node, which turns a workflow into an endpoint other software calls. Metric-based evaluations score the workflow's output over time, and workflow history restores an earlier version when an edit falls short. n8n gives a team one canvas where its automations and its AI agents run as parts of the same workflow.