I’m trying to set up a research system where I make 9 different queries to Perplexity. I’m not really interested in doing something with the actual query result.
Rather, I want to get the Perplexity URL from each and every query and place those into different cells in a Google Sheet.
I can’t find any documentation about Perplexity in make.com documents so I’m reaching out to the community.
Thank you!
1 Like
+1
I am also trying to do something similiar.
Currently I am using prompting to convince the LLM to put the URLs into a super simple JSON structure.
{
“page1”: “https://www.example.com/page1”,
“page2”: “https://www.example.com/page2”,
“page3”: “https://www.example.com/page3”
}
I then send the output of the message content to a JSON “Tool” module so I can separate them out.
The issue I have is that the LLM sometimes, not always will add some sentence like “here are the URLs requested” outside of the JSON structure.
I did just see that chatGPT is introducing an upgraded" Structured Outputs in the API".
hopefully that would be available here soon ???