Zay-M3

Zay-M3 / NaturalSQL

Public

Este es un proyecto para experimentar sobre el conceto de RAG y como se implementa esta tactica de contexto en los modelos grandes de IA (LLM)

14
3
100% credibility
Found Mar 24, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

NaturalSQL is a lightweight Python library that extracts a SQL database schema, indexes it semantically in a vector store, and retrieves relevant tables to provide context for LLMs generating SQL from natural language queries.

How It Works

1
💡 Hear about NaturalSQL

You want to chat with your database in plain English, without memorizing tricky commands, and find this handy tool that makes it possible.

2
🛠️ Get the tool ready

You add the tool to your computer setup with a simple download, picking options for your type of database if needed.

3
🔗 Connect your database

You share the address of your database, like telling a friend where your filing cabinet is kept.

4
✨ Create a smart table guide

The tool peeks inside your database, learns all about your tables and columns, and builds a quick smart index to remember them perfectly.

5
❓ Ask in everyday words

You type a natural question, such as 'show sales from last month' or 'find new customers'.

6
🔍 Get matching tables

It searches its smart guide and pulls up exactly the right tables that fit your question, super fast after the first time.

7
🤖 Let AI craft the query

You hand those tables to your favorite AI chat buddy, who whips up the perfect search command just for your words.

🎉 See your data shine

Run the ready-made command, and voila—your answers appear clearly, making database chats feel effortless and fun.

Sign up to see the full architecture

6 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is NaturalSQL?

NaturalSQL is a lightweight Python library that extracts your SQL database schema—PostgreSQL, MySQL, SQL Server, or SQLite—and builds a vector store with ChromaDB for semantic search. It gives LLMs precise contexto from relevant tables to generate SQL queries from natural language, solving the bloat of heavy RAG tools for small projects. Pip install with DB extras, build the vector DB once, then search queries like "ventas del ultimo mes" in milliseconds thanks to caching.

Why is it gaining traction?

This proyecto ambicioso fue muy fácil de lograr: automatic schema vectorization with a tiny SentenceTransformer model, plus helper prompts for Groq or any LLM, keeps it under 10 deps base. Devs dig it for ~10-15ms cached searches after 2-5s warmup, forced resets, and no-frills setup—ideal over GitHub Copilot for DB-specific NL-to-SQL. Multi-DB support via extras like [postgresql] makes swapping engines painless.

Who should use this?

Backend devs building quick NL query tools for dashboards or internal apps, tired of verbose frameworks. Indie hackers prototyping RAG on grandes LLMs with SQL backends, or teams experimentar con contexto in este proyecto de vida simple. Skip if you need query execution or advanced chaining.

Verdict

Promising for experiments at 14 stars and 1.0% credibility score, with bilingual docs and PyPI ready—but too green for prod without tests or broader adoption. Try it este proyecto en ingles for lightweight NL-to-SQL wins.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.