EduMap — Audio to Mindmap with LLM-Powered Reasoning

EduMap is an AI-powered tool that transforms lecture audio into structured, visual mindmaps.

Website

Screenshots

How It Works

1. Transcribe Audio: Uses Assembly AI to convert uploaded or recorded MP3/WAV into text
2. Extract Concepts: A DSPy MindmapExtractor module identifies central topics & subtopics
3. Generate Mindmap: JSON structure is turned into a mindmap using Graphviz
4. Render & Display: Visual shown in real-time via Streamlit

Technologies Used

Layer	Tool/Library
Transcription	Assembly AI
LLM Based Extraction	Gemini, DSPy
Visualization	Graphviz (graphviz.Digraph)
UI	Streamlit
API Server	FastAPI (for extraction endpoint) + Render (Deployment)

Installation

1. Clone this repo

git clone https://github.com/its-bhavya/edumap.git
cd edumap

2. Install Dependencies

 pip install -r requirements.txt

3. Install Graphviz

# Ubuntu/Debian
sudo apt install graphviz

# macOS
brew install graphviz

# Windows (use choco or installer from official site - https://graphviz.org/download/)
choco install graphviz

4. Setup Environment Variable with Gemini API Key

Create a .env file:

GOOGLE_API_KEY=your_gemini_key
ASSEMBLY_API_KEY=your_assembly_api_key

5. Run the App

a. Start the FastAPI backend (DSPy Extractor):

uvicorn backend.main:app --reload

b. Start the Streamlit frontend:

streamlit run frontend/main.py