Hello Make Community,
Sorry if this has been answered before; I am a noob.
I’m working on a workflow that integrates ChatGPT to generate content for my blog.
I want to pass the output from GPT to create a CSV to upload content to my CMS in Framer easily.
My goal is to maintain consistency across titles, content, and other elements without making multiple calls to GPT for each item.
Initially, I wasted many hours trying to parse the output from ChatGPT, even when provided with an example JSON.
To resolve this, I added the “Create JSON” step in my scenario, which seems to have improved the workflow.
However, I’m still struggling to determine the ideal format or structure that ChatGPT should output to make the JSON creation step smoother and more efficient.
My question:
What is the best output format or structure I should aim for from ChatGPT to ensure seamless integration with the “Create JSON” module in Make? Any tips or best practices would be greatly appreciated.
For context, I’ve included an image of my current workflow setup, where I’m trying to parse data from GPT into a JSON format.
Thanks in advance for your help!
Please note: I created a redundant step with Google Sheets to easily log and inspect the raw output from GPT, which has helped me build the necessary data structures.