Welcome to Ask Rezzy, the β¨ intelligent GenAI assistant built to make your med-school grind easier, faster, and a lot more fun π§ π¬.
Think of Rezzy as your resident genius friend whoβs always ready to:
- Pull out flashcards from memory π
- Recall tricky past-year questions π§Ύ
- Chat with you in real-time π§ββοΈ
- Answer doubts like a walking textbook ππ‘
When you're buried in anatomy books π, flipping between chapters, and stressing over βwhat to revise,β Rezzy becomes your:
- π Search wizard for semantically similar questions
- π§ LLM assistant trained to fetch, retrieve, and reason
- π¬ Live chatbot that streams answers word-by-word
- πͺ Flashcard flipper with perfect timing
βGive me 5 questions on the skeletal systemβ β Rezzy dives into its brain π§ and streams back smartly chosen questions π§Ύβ‘
π Live Demo on Loom
π¨ Figma UI Design
π Drive Link with Dataset
| πΌ Feature | π¬ Description |
|---|---|
| π Semantic Vector Search | Fetches smart matches based on meaning, not just keywords |
| π§ RAG (Retrieval + Generation) | Finds useful content and builds contextual responses |
| π¬ LLM Chat Interface | Streams responses token-by-token like magic πͺ |
| π§Ύ Flashcards & PYQs | Answerable flashcards and previous-year questions |
| π§ββοΈ Study Mode UX | Typing indicators, smooth animations, real-time feel |
| βοΈ Layer | π οΈ Tech Used |
|---|---|
| π± Mobile | React Native + Expo (SDK 50+) |
| β¨ Language | TypeScript (strict mode π§ ) |
| π§ Navigation | expo-router |
| π Backend | Express.js (REST + SSE + WebSockets) |
| π Vector DB | Pinecone / Weaviate / Chroma (you choose!) |
| π§ LLM | OpenAI (Embeddings + Function Calling) |
| β‘ Real-time | WebSockets (Socket.io) + SSE |
| π State | React Query (TanStack) |
| π§ͺ Testing | Jest (for core pipeline) |
interface Question {
id: string;
topic: string;
subtopic: string;
question: string;
type: "multiple_choice" | "true_false" | "short_answer";
options?: string[];
correct_answer: string;
difficulty: "easy" | "medium" | "hard";
tags: string[];
}
interface Flashcard {
id: string;
topic: string;
subtopic: string;
front: string;
back: string;
difficulty: "easy" | "medium" | "hard";
tags: string[];
}