Created an app that connect OpenAI to my production postgres DB using langchain.<p>App: https://askmydb.vercel.app/<p>Check out the demo: https://www.loom.com/share/108e23873cfc41edbb8923cc86edccb3<p>Built using Next JS, Vercel, AWS Lambda and langchain. Everything is stored locally on the user's computer.
How does it know how to extract the related data? I could imagine how it could work with fairly popular "user" table with popular fields (from your loom video example). But it's not clear to me how it might work with more business specific data and tables/fields.
+ What's the full prompt used?