It’s great to see that assistants can use tools to read documents and run Python code without requiring user action.
Since we can upload files, it would also be useful to embed code that tells the assistant what to do with these files.
Currently I created an assistant that uses an function calling to obtain documentation for a specific API and another function calling to check whether a hardware is supported in the application. This is done in a more accurate, faster and cheaper way than the knowledge retrieval.
They work great, but if I want to distribute this feature, I will have to use my computer as a server (or distribute updated data packages to be read in the end user’s computer).
Do I really need a server to distribute these assistants so that they are usable Or is there some other solution that I’m not aware of?