AI Quiz Generator

A Python app that turns any study material — pasted text, a PDF, or a URL — into a polished quiz using the Anthropic Claude API. Includes a Streamlit web UI and a CLI.

---

Features

• Three input types: paste raw text, upload a PDF, or scrape a web page
• Three question types: multiple choice, true/false, short answer
• Configurable: number of questions (1–20), difficulty (easy / medium / hard), mix of question types
• Interactive quiz UI: one question at a time, progress bar, instant scoring
• Export: download results as JSON or CSV (Google Forms-compatible)
• CLI: scriptable, pipeline-friendly, supports all the same options

---

Sample Output

{
  "question": "What is the primary function of mitochondria in a cell?",
  "type": "multiple_choice",
  "options": [
    "A) Protein synthesis",
    "B) Energy production (ATP)",
    "C) DNA replication",
    "D) Waste removal"
  ],
  "correct_answer": "B",
  "explanation": "Mitochondria are known as the powerhouse of the cell because they produce ATP through cellular respiration.",
  "difficulty": "medium",
  "topic": "Cell Biology"
}

---

Setup

1. Clone and install dependencies

git clone https://github.com/your-username/ai-quiz-generator.git
cd ai-quiz-generator
pip install -r requirements.txt

2. Add your Anthropic API key

cp .env.example .env
# Edit .env and replace the placeholder with your real key

Your .env file should look like:

ANTHROPIC_API_KEY=sk-ant-...

Get a key at console.anthropic.com.

---

Usage

Web UI

streamlit run app.py

Opens at http://localhost:8501. Choose your input type, paste/upload your material, configure the quiz settings in the sidebar, and click Generate Quiz.

CLI

# From a text string
python main.py --text "Mitochondria are membrane-bound organelles..." -n 5 -d easy

# From a PDF
python main.py --pdf lecture_notes.pdf -n 10 -d medium --export-json quiz.json

# From a URL, mixed question types, export to CSV
python main.py --url https://en.wikipedia.org/wiki/Photosynthesis \
  --types multiple_choice true_false \
  -n 8 -d hard \
  --export-csv quiz.csv

CLI options

Flag	Description	Default
--text TEXT	Raw study material (quoted string)	—
--pdf FILE	Path to a PDF file	—
--url URL	Web page to scrape	—
-n N	Number of questions	10
-d LEVEL	Difficulty: easy, medium, hard	medium
-t TYPE …	Question type(s) to cycle through	multiple_choice
--export-json FILE	Save quiz as JSON	—
--export-csv FILE	Save quiz as CSV	—

---

Project Structure

ai-quiz-generator/
├── app.py               # Streamlit web UI
├── main.py              # CLI entry point
├── generator.py         # Claude API quiz generation
├── parser.py            # Text / PDF / URL input parsing
├── exporter.py          # JSON and CSV export
├── data/
│   └── sample_material.txt
├── tests/
│   ├── test_parser.py
│   ├── test_generator.py
│   └── test_exporter.py
├── requirements.txt
├── .env.example
└── .gitignore

---

Running Tests

pytest tests/ -p no:playwright

62 tests covering all parser, generator, and exporter functions with mocked API calls — no real API key needed for the test suite.

---

Tech Stack

Layer	Library
AI	Anthropic SDK · claude-sonnet-4-6
Web UI	Streamlit
PDF parsing	PyMuPDF
Web scraping	requests + BeautifulSoup4
Testing	pytest

---

Deployment (Streamlit Cloud)

1. Push this repo to GitHub.
2. Go to share.streamlit.io and connect your repo.
3. In Advanced settings → Secrets, add:

   ANTHROPIC_API_KEY = "sk-ant-..."

4. Click Deploy.