I was creating a custom gpt for a collaborative research project, but OpenAI restricted how many documents I could upload but was generous in the overall file size. So I combined the Arxiv documents we were using and wanted to chat with into one massive pdf and uploaded that.
To make the Python script available to other team members, I deployed it as an Azure function to a shared development environment where they could access and test the endpoint in real time. This allowed them to run it for future gpt’s without requiring local setup.