Last year, I developed a CLI application called FerriteChatter using Rust to utilize the OpenAI API for leveraging LLMs.

When creating an application that acts as a client for external APIs, its behavior and implementation might change based on the responses received.

In this presentation, I will showcase a case where we used the OpenAI API as an example to execute the API at build time and generated Enums and implementations using proc_macros based on the responses.

Additionally, I will address the risks associated with using GitHub Actions when adopting such an implementation strategy.