AI Quiz Generator in 2026: Build a Free Self-Hosted Quiz App with Anki (30k Stars) + LangChain - 5 Working Patterns

Anki (29,543 stars) stores the questions, LangChain (143,470) generates them from any source material, and an LLM keeps difficulty calibrated - a free quiz system for teachers, tutors and self-learners.

💡 What You Will Learn

Anki (29,543 stars) stores the questions, LangChain (143,470) generates them from any source material, and an LLM keeps difficulty calibrated - a free quiz system for teachers, tutors and self-learner

## The short answer The fastest way to build a quiz generator: use an LLM to create questions from your source material (lecture notes, textbook chapters, articles), store them in Anki (29,543 stars) for spaced repetition, and auto-grade with the same model. Everything is free and self-hosted. ## Pattern 1 - Generate questions from any text ```python import openai text = open("chapter3.txt").read() resp = openai.chat.completions.create( model="gpt-4o-mini", messages=[{"role": "user", "content": "Create 10 multiple-choice questions from this text, 4 options each, " "with correct answers. Return JSON. " + text[:8000]}]) print(resp.choices[0].message.content) ``` ## Pattern 2 - Cloze deletion for memorization Ask for fill-in-the-blank cards ("The capital of France is ___") - these are the most effective for spaced repetition in Anki. ## Pattern 3 - Auto-grading ```python resp = openai.chat.completions.create( model="gpt-4o-mini", messages=[{"role": "user", "content": f"Grade this answer to '{question}' out of 10 with one-line feedback: {answer}"}]) ``` ## Pattern 4 - Difficulty calibration Have the model tag each question 1-5 by difficulty, then schedule harder questions less frequently - Anki handles the scheduling. ## Pattern 5 - Import to Anki Export questions as CSV (`Front,Back,Tags`) and import into Anki, or use AnkiConnect for scripted sync. ## Real numbers - Generating 100 questions from a textbook chapter costs ~$0.05-0.2 with a mini-class model. - Anki's spaced repetition lifts 30-day retention from ~30% (cramming) to 80%+. - The full stack runs on a laptop - no cloud accounts needed. ## FAQ **Q: Are the questions any good?** A: Mini-class models generate good MCQs from clear source text; review and edit the first batch, then calibrate your prompt. **Q: Can students use it too?** A: Yes - feed it your notes, generate self-quizzes, and Anki handles review scheduling. **Q: What about math questions?** A: LLMs handle basic arithmetic and algebra; for LaTeX-heavy questions, ask for LaTeX output and render with MathJax in Anki.