Assistance Needed for Querying Multiple Databases with OpenAI API

I have a bot that needs to interact with various databases, including Postgres, Snowflake, Oracle, BigQuery, MariaDB, and local CSV/JSON files. Similar to the pandas’ agent, I can provide an N number(supposed 3 databases and 1 CSV file) of data frames and query any data frame (CSV) as needed. However, using a “SELECT * from table” query and saving this data into CSV, and giving that CSV to any existing agent for every user question when accessing databases is not scalable.

I have explored approaches such as pandas agents and langchain SQL agents, but scalability becomes a major issue when users query data from multiple databases.

Could you please suggest any pipelines, approaches, or tools that could help address this use-case?
Your insights and suggestions would be greatly appreciated.
Thank you!

You can explore,

if it helps!