So, I want to dedicate some of my free time to learning this, but I’m unsure if it will work or if it’s a good idea, since my experience with APIs or web-dev is zero, but I want to learn !
I aim to create a chatbot that aids in creating scripts from software I use. The software has a RAD plugin that lets you quickly create python scripts for automation. It comes with a library.
I have all the software API documentation in .chm format and all the necessary library files.
What I want:
Develop a custom GPT (not an assistant).
Currently, it works quite well, but it hallucinates classes and methods.
Let’s say we add an instruction saying if the user types /searchapidocs:
The chat analyzes the script request and asks for the correct methods/classes from the API endpoint.
It checks against the embedded library (knowledge base or another call to a vector database?).
Then, the chat retrieves all the required methods/classes from the API documentation, cross check with the library, and provides the best answer.
Is this possible with Custom GPTs?
If so where do I start ? If someone can give me some sort of brief bullet point list on how to achieve this, I can do deeper research myself.