Core Purpose:

A persistent RAG (Retrieval-Augmented Generation) system that answers questions about your documents using local AI models with guaranteed database persistence.

This application can ingest data from multiple sources:

• A text file (facts.txt)
• a DuckDB database (ducks.duckdb)
• a SQLite database (places.sqlite)

Running with Docker

Build the Docker Image

docker build -t rag-app .

Run the Application

This will start a web server on port 8501.

docker run -p 8501:8501 rag-app

Open your browser and navigate to http://localhost:8501. Use the sidebar to select the data source you want to chat with.