🎓 Study-Pilot

Study-Pilot is an AI-powered study assistant that explains concepts in different personalities (modes), generates flashcards, and optionally converts explanations into audio.
Built for fast learning, engagement, and experimentation with AI personalities.

---

✨ Features

• 🧠 AI Study Assistant

  • Clear, beginner-friendly explanations
  • Step-by-step reasoning

• 🎭 Multiple Personality Modes

  • study-buddy
  • mr-broke
  • captain-quizhook
  • gym-bro
  • detective-chalk
  • sleep-deprived-grad
  • customer-support
  • unpaid-intern

• 🔊 Text-to-Speech (Optional)

  • Converts AI responses into audio via a private Hugging Face TTS Space
  • Audio served statically via FastAPI

• 🧾 Conversation Memory

  • PostgreSQL-backed storage
  • Persistent conversations and modes

• 🗂️ Flashcard Generation

  • Generate summaries and flashcards from conversations

• 🌐 Frontend

  • React + Vite + TailwindCSS

---

🏗️ Project Structure

study-pilot/
│
├── backend/
│   ├── main.py              # FastAPI app
│   ├── main_ai.py           # AI-only testing entry
│   ├── llm_utils.py         # Prompt building, modes, TTS logic
│   ├── prompts.py           # System + mode prompts
│   ├── database.py          # DB connection
│   ├── database_models.py   # SQLAlchemy models
│   ├── models.py            # Pydantic schemas
│   └── init.py              # DB initialization
│
├── frontend/
│   ├── index.html
│   ├── src/
│   ├── package.json
│   └── vite.config.js
│
└── README.md

---

⚙️ Tech Stack

Backend

• FastAPI
• OpenAI API (gpt-4o-mini)
• PostgreSQL
• SQLAlchemy
• Hugging Face Spaces (TTS via Gradio Client)

Frontend

• React
• Vite
• TailwindCSS

---

🚀 Getting Started

1️⃣ Clone the repository

git clone <your-repo-url>
cd study-pilot

---

2️⃣ Backend Setup

Install dependencies

pip install fastapi uvicorn sqlalchemy psycopg2 openai gradio-client

Configure Database

Update database.py with your PostgreSQL credentials:

db_url = "postgresql://USER:PASSWORD@localhost:5432/FastAPI"

Create DB tables

python init.py

---

3️⃣ OpenAI API Key

Set your OpenAI key (recommended via environment variable):

export OPENAI_API_KEY="your-key-here"

⚠️ Do NOT hardcode API keys in production.

---

4️⃣ Run Backend

uvicorn main:app --reload

Backend:

http://localhost:8000

---

5️⃣ Frontend Setup

cd frontend
npm install
npm run dev

Frontend:

http://localhost:5173

---

🔊 Text-to-Speech (Optional)

• Uses a private Hugging Face Space
• Audio generated via gradio_client
• Saved to:

  static/audio/
  

• Served at:

  /static/audio/<filename>.wav
  

⚠️ Depends on GPU availability and Space uptime.

---

🎭 Modes Explained

Mode	Description
study-buddy	Friendly and patient
mr-broke	Dramatic, money metaphors
captain-quizhook	Pirate-themed
gym-bro	High-energy motivation
detective-chalk	Noir detective
sleep-deprived-grad	Chaotic but smart
customer-support	Calm and professional
unpaid-intern	Bitter but competent

Modes influence prompt style, temperature, and penalties.

---

🧪 Example API Call

POST /study

{
  "user_text": "What is RAM?",
  "mode": "gym-bro"
}

Response

{
  "response": "RAM is like your brain during a workout...",
  "audio_file_path": "/static/audio/example.wav"
}

---

🛠️ Known Limitations

• Hugging Face GPU quotas
• Cold-start latency
• Audio length may truncate based on model behavior

---

📌 Future Improvements

• Streaming audio
• More voices
• Auth + user profiles
• Docker deployment

---

🏁 Built For

• Hackathons
• Learning tools
• AI experimentation
• Educational demos

---

🧠 Final Note

If it teaches you and makes you laugh — it’s working.