👨‍💻👩‍💻 QuickKanban

QuickKanban is a real-time, collaborative kanban board with an AI Copilot assistant for intelligent task tracking, built for efficient project planning and team coordination.

---

🚀 Tech Stack

AI Copilot (LLM Integration)

QuickKanban includes a real-time AI Copilot that delivers task insights as your board updates using a Retrieval-Augmented-Generation (RAG) pipeline. All queries first go through the Express.js backend API.

• Embeds task updates using Gemini LLM embeddings
• Stores and retrieves vectors with Qdrant
• AI summaries, suggestions, and insights are streamed to user within ~500ms
• A contextual thread is maintained between Gemini and user using Redis
• Gemini and Qdrant interactions are served through dedicated FastAPI microservices

As users add more tasks, the AI Copilot can provide deeper insights for the expanding board - enabling dynamic reasoning and memory.

Features

• ⚡ Real-time updates using WebSockets (Socket.IO)
• 🤖 AI Copilot Assistant that summarizes and reasons about tasks in real time
• 🧠 RAG-based vector search powered by Gemini + Qdrant
• 🧵 LLM memory and thread context via Redis
• 👥 Live user presence tracking in collaborative kanban rooms
• 🖱️ Live cursor indicators to show teammates' movements in real-time
• 📌 Add tasks fluidly to boards; drag-and-drop with column transitions
• 💾 Persistent task storage with MongoDB

System Architecture

• Frontend: React + Socket.IO for live collaboration and presence
• Main API: Node.js + Express + WebSockets
• AI Microservices: FastAPI-based embedder, retriever, and query services
• Gemini + Qdrant: RAG backend with embeddings + vector search
• Redis: Stores temporary conversational thread between the LLM and user and serves as a general cache for board
• MongoDB: Durable task + board storage

---

Setup & Deployment

QuickKanban is fully containerized and deployed using AWS.

Containerization

• Each service (backend, AI microservices, Redis, Nginx) runs in its own Docker container
• Containers are managed using ECS with EC2 launch type

AWS Deployment

• Amazon ECS (EC2 Launch Type): Orchestrates microservices
• Amazon EC2: Hosts all running containers and mounts SSL certs
• Amazon ECR: Stores Docker images
• Nginx: Handles HTTPS and reverse proxies to backend

Deployment files:

• ecs-task-definition.json
• nginx.conf
• Dockerfile (for both backend and Nginx)

---

🤝 Contributing

Please feel free to contribute! This is a fun, ambitious project with lots of room to grow.

🛠️ Local Development Setup

1. Fork this repository
2. Clone your fork:

git clone https://github.com/mhit30/quickkanban.git

Note: Make sure Qdrant, MongoDB, Redis running on your system. Also obtain a Gemini API Key from Google GenAI to connect to Gemini.

 cd backend && npm install

 cd frontend && npm install

For backend (with Nodemon):

  nodemon index.js

For frontend (with React Vite)

  npm run dev

🔄 Making PRs

• Use feature branches:
  git checkout -b feature/your-feature-name

• Keep pull requests scoped and descriptive:

  • What feature or fix you made
  • Why its helpful

📄 License

This project is licensed under the MIT License.
You're free to use, modify, and distribute this software with proper attribution.