Why we built it When working on various projects at deepsense.ai (we're part of the org), we often needed a way to fetch data from databases using natural language queries. The traditional text-to-SQL approach was powerful but failed at understanding domain-specific queries and usually yielded inconsistent results. So, we built db-ally to streamline this process and simplify data retrieval with natural language queries. By defining specific use cases, db-ally makes querying efficient, predictable, and easy to manage.

Asking for feedback As this is an R&D project, we’re keen to hear your thoughts and feedback. Give db-ally a try and let us know how it works for you. How are you currently handling natural language queries to your databases? What challenges have you faced?

You can find the documentation and repo on GitHub: https://github.com/deepsense-ai/db-ally

We’re looking forward to your insights on what would be most useful for you as we develop it further to meet your needs.

Looking forward to your feedback.