SQLRAG with Gemini

SQLRAG with Gemini is a Streamlit-based demo for asking natural-language questions over a SQLite database and generating SQL-assisted answers with Gemini through LangChain.

Features

• Natural-language to SQL workflow
• SQLite database support
• Gemini integration through langchain-google-genai
• Streamlit user interface
• Simple local setup for experimentation

Requirements

• Python 3.10+
• Google Gemini API access
• A local SQLite database

Setup

git clone https://github.com/emredeveloper/SQLRAG-with-Gemini.git
cd SQLRAG-with-Gemini
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt

Create a .env file in the project root:

GOOGLE_API_KEY=your_gemini_key

Run

streamlit run main.py

Suggested Improvements

• Add sample database schema and seed data
• Add SQL safety checks before execution
• Add examples for common questions
• Add tests for query generation and database helpers

License

MIT