n8n Assistant Turns Plain-Language Requests Into Runnable Workflows

n8n has launched n8n Assistant, an AI-powered way to turn a plain-language automation request into a working workflow. As n8n Blog reports, the assistant plans the node graph, places it on the canvas, requests credentials when needed, runs the workflow, and helps fix failures.

That distinction matters for teams building with AI. Instead of producing an opaque script or a one-off result, n8n Assistant leaves behind a standard n8n workflow: visible, editable, logged, and available for someone else to maintain. Users can inspect each node and execution, change routing or filters, and add approval steps before activating an automation.

The assistant also asks follow-up questions when a request is unclear and requires explicit confirmation for credential access and workflow activation. In n8n’s example, it builds an eight-node process for triaging form submissions, discovers an enrichment edge case, applies a fix, and reruns the workflow.

Usage draws from each plan’s AI credit allocation, with more complex debugging consuming more credits. The allocation is separate from the previous AI Workflow Builder, which this product replaces.

On n8n Cloud, the feature is enabled by default for new instances, except Enterprise Cloud instances. Docker-based self-hosting requires user-provided keys, additional environment variables, and n8n version 2.36 or later. npm self-hosting is unsupported, while an Enterprise release is planned. The feature remains behind a preview flag, and production review, third-party setup, and credential creation still belong to the user.

Source: n8n Blog