RAGnarok

Fast and modern RAG powered research and documentation assistant for professionals

Tech Stack

Overview

The project consists of four main components:

• Ingestion Pipeline
  Reads documents (PDFs, text, image and audio transcriptions), splits and vectorizes them, and stores embeddings in a vector database.

• Backend API
  A FastAPI service exposing endpoints for querying the vector store, invoking SoTa LLMs such as Deepseek, and streaming chat responses.

• Frontend UI
  A NextJS + Typescript chat interface that calls the Backend API for conversational RAG.

• Deployment Manifests
  Dockerfiles and Docker Compose manifests to deploy entire project locally or in the cloud.

Prerequisites

• Docker & Docker Compose
• Node.js (v20+) and Bun
• Python (v3.11+)
• LLM API key (Groq, Gemini, etc.)
• AstraDB Vector database credentials

Project Structure

.
├───.github
│   └───workflows
├───assets
├───backend
│   └───app
│       ├───models
│       ├───routes
│       ├───services
│       └───utils
├───deployment
├───frontend
│   ├───public
│   └───src
│       ├───app
│       │   └───test
│       ├───components
│       │   └───ui
│       ├───context
│       └───lib
└───ingestion-pipeline

20 directories, 65 files

Setup

1. Clone the repo

   git clone https://github.com/3xCaffeine/rag.git
   cd rag

2. Configure environment variables

   Copy the example .env.example into each service folder and add the API keys, URLs and AstraDB database credentials.

3. Run the backend API and Ollama embedding model with Docker Compose

   docker-compose up -d
   

Note

In case GPU support is unavailable on the machine, utilize a hosted text embedding model (OpenAI Ada, Jina AI, Cohere etc.) and remove the Ollama section from the Compose config.

4. Run the dev server for the chat frontend

   bun i
   bun dev

5. (Optional) For remote deployment with custom domains, configure the deployment/Caddyfile otherwise remove the Caddy section from Compose configuration.

Expected Result

$ docker compose ps
NAME            IMAGE                  COMMAND                  SERVICE   CREATED         STATUS                   PORTS
backend-api-1   ragbackend:latest      "fastapi run app/mai…"   api       3 minutes ago   Up 3 minutes (healthy)   0.0.0.0:8000->8000/tcp
ollama          ollama/ollama:latest   "/bin/sh ./run_model…"   ollama    9 seconds ago   Up 9 seconds             0.0.0.0:11434->11434/tcp

Usage

Visit Ragnarok chat at http://localhost:3000 in the browser.

Demo

Authors

This project was built for the AISoC Chronos Hackathon 2025 by the 3xCaffeine team.

• Sourasish Basu (@SourasishBasu)
• Swapnil Dutta (@rycerzes)
• Vaibhav Singh (@monkeplication)
• Shourya Merchant (@dotpixel123)

Version

Version	Date	Comments
1.0	May 14th, 2025	Revised release