Hi everyone,
I’m looking to build a multi-step content automation and wanted to get some advice on the logic—specifically regarding the internal linking component.
I want to automate the flow from a finished draft to a published post while using Claude (Anthropic) to handle the heavy lifting for SEO.
Hi @ivyanna, welcome to the n8n community!
The answer depends on how you structure control and validation in your workflow.
First of all, I would not let Claude directly modify and publish the document in a single step. The key is to separate generation from validation. Let the model suggest the internal links, but always verify them before writing anything back to Google Docs.
Hey welcome, honestly the easiest approach here is to keep a Google Sheet as your internal link database (URL + anchor text pairs) and pass that into Claude’s prompt along with the draft content from Google Docs, let Claude insert the links contextually. There’s actually a template on n8n.io for almost exactly this if you search “WordPress SEO Content Publisher with Anthropic AI”.
Hi @ivyanna Welcome to the community!
Your goal is nice! But your approach to this problem should be different as WHERE you are facing the problem and let us help you in that, although for the best approach just DO NOT use AI agents everywhere in production or even mid scale use cases they break, consider understanding n8n’s core data flow logics and only use AI agent when you actually need LOGIC and REASONING else if you are using AI to just call sheets or get data from sheets then that would really not be good, and also consider using OpenAI models as they are pretty much really works all the time with tool calling.