QueryBox

A RAG-powered chatbot that retrieves relevant context from diverse sources using a vector database and generates precise, context-aware answers with an LLM.

---

Demo

• Watch the demo recording: https://www.loom.com/share/ea8a35c957784b0280f09998dfe10346

• Screenshots of the app are available in the img folder.

Features

• Upload multiple documents: Select and query any uploaded document.
• Natural Language to SQL: Hybrid RAG approach for converting questions to SQL.
• Database integration: Connect your database; QueryBox extracts schemas and retrieves context for LLM-generated SQL queries.
• Context-aware answers: LLM processes SQL results and returns precise responses.

---

In Progress

• Hybrid RAG Natural Language SQL Query:
  Connect your database, QueryBox extracts relevant schemas, LLM generates SQL queries, and the results are fed back into LLM to produce accurate, context-aware responses.

---

Folder Structure

/frontend   # Next.js app
/backend    # FastAPI app

---

Docker Setup

QueryBox uses Docker Compose to run the frontend, backend, and required services (PostgreSQL, Redis, Qdrant) together.

Requirements

• Docker (>= 24.x)
• Docker Compose (included with Docker Desktop)

Environment setup

Before running Docker, copy each directory's .env.example to .env and set required variables.

Commands (run from repo root):

cp backend/.env.example backend/.env
cp frontend/.env.example frontend/.env

Build and Run

From the root of the repository:

1. Build all services: docker-compose build

2. Start all services in detached mode: docker-compose up -d

3. Verify running containers: docker ps

You should see containers for:

• frontend-app (Next.js): http://localhost:3000
• backend-api (FastAPI): http://localhost:8000
• postgres-db, redis, qdrant

---

Stop and Remove

• Stop all containers: docker-compose down

• Stop and remove containers, networks, and volumes: docker-compose down -v

---

Notes

• The frontend communicates with the backend using the Docker network: use http://backend:8000 inside Docker.
• Update .env files in /backend for database credentials, Redis, and Qdrant configuration.
• The backend Dockerfile is reused; the frontend Dockerfile is in /frontend/Dockerfile.

---

License

MIT License