Latest
What n8n's Webhook node actually does
n8n's Webhook node sits at the entry point of a workflow: an external service sends an HTTP request, and the node catches it and starts execution. According to n8n's official documentation, the node supports the standard HTTP methods — DELETE, GET, HEAD, PATCH, POST, and PUT — and accepts payloads up to 16 MB (adjustable on self-hosted deployments via the environment variable `N8N_PAYLOAD_SIZE_MAX`).
What n8n's Webhook node actually does
n8n's Webhook node sits at the entry point of a workflow: an external service sends an HTTP request, and the node catches it and starts execution. According to n8n's official documentation, the node supports the standard HTTP methods — DELETE, GET, HEAD, PATCH, POST, and PUT — and accepts payloads up to 16 MB (adjustable on self-hosted deployments via the environment variable N8N_PAYLOAD_SIZE_MAX).
The node is particularly useful when a service has no dedicated n8n trigger node. It can also work in reverse: by choosing the "When Last Node Finishes" response mode, the webhook waits for the workflow to complete and returns the final output to the caller — effectively making the workflow behave like an API endpoint, as n8n describes it. A third mode, "Using 'Respond to Webhook' Node", lets a dedicated node inside the workflow control exactly what gets sent back. A streaming response option is also listed, intended for real-time output from nodes that support it, such as the AI agent node.
Authentication is optional but available: Basic auth, Header auth, or JWT auth can be required from any caller. The URL path is randomly generated by default to avoid conflicts, but can be set manually — useful when prototyping a consistent API surface. n8n maintains separate test and production webhook URLs; the test URL activates only when you explicitly select "Listen for Test Event" or run the workflow manually, while the production URL becomes active when the workflow is published.
What to try now — and what to verify before committing
The test/production URL split makes initial experimentation low-risk: you can send real events to the test URL during development without touching an active production workflow. That separation is a reasonable starting point for any team evaluating whether the node fits a given integration.
Several questions fall outside what the official documentation alone can answer: how the node performs under sustained high-volume traffic, whether the 16 MB payload ceiling is a hard constraint in cloud-hosted plans, and how streaming response behaves in practice beyond nodes explicitly listed as supported. These are evaluation questions worth testing in your specific environment, not conclusions that can be drawn from the documentation alone. Chooscope has not independently tested any of these scenarios.
Sources
Change history
- rev.1Published
