Studium 📚

Transform lecture transcripts into engaging study videos, delivered straight to your inbox.

🎯 Inspiration

We noticed that very few classmates were rewatching lecture recordings, even when they struggled with the material. The barrier wasn't a lack of motivation—it was time. Sitting through hours of lecture footage to review key concepts felt overwhelming. We created Studium to transform dense lecture transcripts into engaging, digestible study videos that students would actually want to watch.

✨ What it does

Studium converts lecture transcript PDFs into personalized study videos and emails them directly to students. Users simply:

1. Upload their lecture transcript PDF
2. Enter their email address
3. Receive a custom-generated study video within minutes

The AI analyzes the transcript, generates an engaging script, creates a video explanation, and delivers it via email—turning hours of lecture content into focused study resources.

🏗️ Architecture

Frontend (Framer) → Flask API → n8n Workflow → OpenAI → OpenNote API → Email

Tech Stack

• Frontend: Framer (React-based UI)
• Backend: Flask (Python)
• Workflow: n8n (automation platform)
• AI Processing: OpenAI API (script generation)
• Video Generation: OpenNote API
• Deployment: ngrok (development), ready for production deployment

🚀 Getting Started

Prerequisites

• Python 3.8+
• ngrok (for local development)
• n8n account
• OpenAI API key
• OpenNote API key

Installation

1. Clone the repository

git clone https://github.com/NickB-30/CruzHacks2026.git
cd CruzHacks2026

2. Install Python dependencies

pip install -r requirements.txt

3. Set up environment variables

export N8N_TRANSCRIPT_WEBHOOK="your-n8n-webhook-url"
export TEST_MODE="false"  # Set to "true" for testing without n8n

4. Run the Flask backend

python app.py

5. Expose Flask with ngrok (in a new terminal)

ngrok http 5000

6. Update Framer frontend

• Copy the ngrok URL
• Update the fetch URL in your Framer component to use the ngrok URL

n8n Workflow Setup

1. Create a new workflow in n8n
2. Add a Webhook node (POST method)
3. Configure the following nodes:
   • PDF text extraction
   • OpenAI node (script generation)
   • OpenNote API call (video generation)
   • Email node (delivery)
4. Activate the workflow
5. Copy the production webhook URL to your Flask environment variables

📁 Project Structure

CruzHacks2026/
├── app.py                 # Flask backend API
├── requirements.txt       # Python dependencies
├── uploads/              # Temporary storage for uploaded PDFs
├── README.md             # This file
└── framer/               # Framer frontend code (separate)

🔑 API Endpoints

POST /upload/transcript

Upload a transcript PDF and trigger video generation.

Request:

• file: PDF file (multipart/form-data)
• email: User's email address (string)

Response:

{
  "message": "Transcript processed successfully",
  "filename": "lecture.pdf",
  "status": "success",
  "n8n_response": "..."
}

GET /health

Health check endpoint.

Response:

{
  "status": "healthy"
}

🛠️ Configuration

Environment Variables

• N8N_TRANSCRIPT_WEBHOOK: Your n8n webhook URL
• TEST_MODE: Set to "true" to skip n8n and test file uploads only

File Constraints

• Allowed formats: PDF only
• Max file size: 50MB
• Email validation: Standard email format required

🧪 Testing

Test Mode

Enable test mode to verify file uploads without calling n8n:

export TEST_MODE="true"
python app.py

Manual API Testing

curl -X POST -F "file=@transcript.pdf" -F "email=test@example.com" http://localhost:5000/upload/transcript

🎨 Frontend

The frontend is built with Framer and features:

• Modern, gradient-based UI
• Drag-and-drop file upload
• Real-time status updates
• Responsive design
• Email validation

🚧 Challenges

Our biggest challenge was fully automating the AI pipeline with n8n. As first-time users, we had to learn webhook configuration, data flow management, and error handling—all within the hackathon timeframe. Successfully connecting Flask, n8n, OpenAI, and OpenNote APIs required creative problem-solving and persistent debugging.

🏆 Accomplishments

• Built a complete AI pipeline from scratch in one weekend
• Successfully integrated 4+ different technologies into a cohesive product
• Created an elegant, user-friendly interface
• Solved a real problem that students face daily

📚 What We Learned

• Designing and implementing complex workflows with n8n
• Structuring API communications between multiple services
• Handling asynchronous processes in user-facing applications
• The importance of breaking ambitious ideas into manageable components

🔮 What's Next

• Yuja Integration: One-click study videos directly from lecture pages
• Multiple Formats: Add flashcards, quizzes, and summary notes
• Customization: Allow users to choose video length and focus topics
• Scale: Expand to universities nationwide
• Enhanced AI: Improve script quality and add citations/sources

👥 Team

Built with ❤️ at CruzHacks 2026

📄 License

This project was created for CruzHacks 2026.

🙏 Acknowledgments

• CruzHacks 2026 organizers
• OpenAI for providing powerful language models
• OpenNote for video generation capabilities
• n8n community for workflow automation tools

---

Note: This is a hackathon prototype. API keys and production credentials should be properly secured before deployment.