n8n workflow · content automation

One post in. Every channel out.
A human click in between.

This is a simulation of the exact workflow you download: an article arrives by RSS or webhook, is fetched and cleaned, then three LLM prompts draft a LinkedIn post, a numbered X thread and a newsletter blurb. Nothing publishes until you approve the email — then LinkedIn, Buffer and your newsletter webhook fire in parallel.

19 nodes · real n8n export 3 prompts with editorial guardrails Human approval before anything posts 15-min undo window, configurable
Replays the 654-word sample article shipped in the download, node by node.
Workflow canvasIdle
Drag sideways to see the whole workflow →
Execution log
    Ready. The sample article is loaded — press Run sample article to watch it move through every node, with the real transformed content shown at each step.
Node outputnewest first

This page replays the node logic and the sample outputs that ship in the download (sample-input.json / sample-outputs.md). In a live run the three drafts are generated by your own OpenAI credential, so wording varies run to run — structure, guardrails and the approval flow are exactly what you see here. Publish delays are compressed for the demo.