Getting Structured Responses from LLMs

[@gmcwhinney]

Nicely done 🤩 Going down the prompt route to request structured data like this is so flexible and usable across the different models... Genius!

[@daniel_migizi]

This is awesome. Thanks for the demo.

[@familiekontoen1955]

Great video! Really well explained and easy to follow and adapt as a newbie.

[@charliebailey2755]

This worked well for a single item but there were issues that took me forever to solve with a list of values! Follow a similar process to this but from the response date from GPT, create a new variable with the content, update variable with a join with just a basic line break (shift + enter), then use JSON decode. I found the solution in a help video posted a couple months ago with some issues with Bubble API data - kzbin.info/www/bejne/ZqSWeGxomKpmgbMfeature=shared&t=601 (solution starts at 10 minute mark)

[@MegaPakster]

I have not had any issues with JSON arrays from LLMs using the method from Xano.