HTTP request tool, running non-stop. (Tavily)

Questions
1

Hi guys, Im creating a sales brief generator, where it’ll scrape the net using Tavily.

Version 1: i fix the queries but then let AI compile the output.

I’ve done one using a HTTP request node, where it I manually ‘fix’ the queries and was able to get the right output, but because there many queries/results, i sometimes faced issue where the language model itself timed out.

Version 2: Let AI determine the query and compile the output (video)
However, somehow the AI agent just runs indefinitely, initially I thought it was just because it ran through multiple queries, but then I realised the AI agent node was passing through data to Language model > HTTP request node back to AI agent node and back to language model again, causing it to loop indefinitely? Im not sure if im even understanding the flow correctly. It’s my first time working with HTTP request tools, so Im not sure if im setting it up properly.

Any pointers to help me understand it would be better, as I’m not too sure how the workflow ‘flows through the HTTP request node and the language model through the AI agent node.

Hope you guys understand what im saying, please help if possible. Thanks

2

It sounds like your AI agent might be stuck in a loop because the HTTP Request node is continuously passing data back and forth. To fix this, ensure you set clear exit conditions or limits in your workflow. You can also use the Tavily node directly in n8n for streamlined web scraping, which might simplify your setup. Check out the [docs.n8n.io]( HTTP Request Tool node documentation | n8n Docs ) for more guidance on using the HTTP Request tool effectively.