Show HN: Mnemiq – Open-source text-to-SQL tuned to your own database

[flagged]

Hi everyone! This is Paulina, one of the co-founders of Agentic Fabriq (the team that built and open-sourced mnemiq). I wanted to come on here and share a bit about why we built mnemiq.

Recently we've been working more with data agents on the governance side, and text-to-SQL kept coming up as another major struggle. So we dug into it more and realized that a big gap was most text-to-SQL engines are like a black box - you have no control over the process of how the text you give in becomes SQL. Even though there are many text-to-SQL vendors, and even though they publish high accuracy rates, it's not necessarily going to give you better text-to-SQL on your own data because everyone's data is different.

We built mnemiq as the text-to-SQL engine that you can edit/configure so it works best on your data, not on ours. You can try different models, retrieval, enrichment, etc. to see which setup does the best for you.

If you end up giving it a try, I'd love to hear how it went! It works with Postgres, SQLite, Oracle, Snowflake, and Databricks.