Multiple input datasets in the assistant

API
, , , ,
1

I developed an assistant that integrates more than 9 CSV datasets using the file_id parameter. The idea is to ask questions about these datasets, relating information across them. However, due to the large volume of quantitative and strategic data, the assistant began showing the “rate limit exceeded” message for tokens per minute on the second question.

Considering that I plan to continue adding more datasets, what would be the best alternative for me? Is it still worthwhile to use the assistant, or should I consider another method? I’m evaluating the option of using more than one assistant in the same thread.

2

This is the challenge: I keep thinking about how to avoid being limited by this issue, especially since I have many agents. I realize we are very dependent on a single company. I’m still considering building a decentralized multi-agent framework, but if it evolves as planned, it might lead to issues I can’t control.