Real-time AI chatbot with multimodal capabilities and long-term memory
Built with MERN, Gemini API, and Pinecone Vector DB
Nebula is a full-stack AI chatbot capable of understanding, reasoning, and creating.
It uses short-term and long-term memory (STM + LTM) to retain chat context and knowledge over time, powered by Pinecone Vector DB.
Core Capabilities
- Real-time AI communication via Socket.io
- Context-aware memory using STM + LTM
- Code generation & explanation
- Image understanding and generation via Gemini API
- Cloud media management with Cloudinary
- Optimized MERN architecture with Shadcn UI + Tailwind
Frontend (React + Redux + Tailwind) ↓ Socket.io (real-time bridge) ↓ Backend (Express + Node.js) ↓ AI Layer (Gemini API + OpenAI + Pinecone Memory) ↓ Database (MongoDB) ↓ Cloud Storage (Cloudinary)
- Real-time bi-directional chat (Socket.io)
- Vector-based knowledge retention (Pinecone)
- Image understanding and generation
- Modular service-based backend
⚠️ Note: The current version of Nebula is not yet fully responsive and is optimized for desktop view.
⸻