allow to use custom GPT assistant

[dz0]

Describe the need of your request

I want to have all my codebase in the context.

Proposed solution

OpenAI GPT assistants allow to have RAG based on full codebase (via file-search).

It would be faster and require less tokens than sending (limited) code-context each time.

Additional context

No response

[Mte90]

I am looking for something similar that I can use with custom providers...