self-hosted conversational analytics layer for databases

I've been working on a self-hosted tool that lets everyone query databases using natural language and generates/validates SQL before execution. The concept is that everyone even not being good or knowing SQL to be able to get insights form his/her data. It runs in your own environment and supports Docker, with no requirement to send database rows to a hosted SaaS. As I data engineer built it to free data engineers from getting requests like "compare this month's results with last month" and wanted a more practical interface on top of existing databases. Even the CEO should be able to get results on his own. Community edition currently supports Postgresql and Mysql, semantic metadata, SQL validation, permissions/privacy controls and review-before-execution. It takes more or less 5 minutes to set it up. Docker setup + documentation: github.com/savvina-ai/savvina Drop a star or comments in discussions! I'd love feedback from people running their own databases: would you actually use something like this, and what would it take to put it in your stack?