FeynmannAI

Master complex topics by teaching them to an AI.

FeynmannAI is an interactive learning platform based on the Feynman Technique: the idea that you learn best when you explain a concept simply. Instead of explaining to a person, you teach a curious AI student who asks follow-up questions to test your understanding.

Here is the demo video:

Here is the devpost for the hackathon submission: devpost url

Project Structure

This repository is divided into two main parts:

• Backend: A Python/FastAPI server that handles the AI logic, speech processing (ElevenLabs), and memory management.
• Frontend: A Next.js web application that provides the chat interface, voice recording, and dashboard.

Quick Start

Prerequisites

• Node.js (v18+)
• Python (v3.10+)
• API Keys for:
  • Gemini (Google AI)
  • ElevenLabs (Voice)
  • Supabase (Auth & Database)

1. Backend Setup

Navigate to the backend directory and follow the detailed instructions.

cd backend
python -m venv env
source env/bin/activate  # or .\env\Scripts\activate on Windows
pip install -r requirements.txt
uvicorn main:app --reload

2. Frontend Setup

Navigate to the frontend directory and follow the detailed instructions.

cd frontend
npm install
npm run dev

Open http://localhost:3000 to start learning!

Tech Stack

• AI Core: Google Gemini Flash (Vision & Text)
• Voice: ElevenLabs (Scribe v2 & TTS)
• Backend: FastAPI (Python)
• Frontend: Next.js (React) + TailwindCSS
• Database/Auth: Supabase
• Rendering: ReactMarkdown + KaTeX (for math equations)