Make your notes quiz you, because you won't

What it does

Examor is a self-hosted web application that ingests your documents and automatically generates questions based on their contents. An LLM grades your responses against the source material and schedules follow-up reviews using a simplified Ebbinghaus forgetting curve. It is essentially a spaced-repetition system where the quizzes are built and scored by AI.

The interesting bit

The twist is the feedback loop: your exam score directly stretches or shrinks the interval until the next review, and the entire spaced-repetition scheduler is reportedly just eight lines of code. The project also includes a community question bank seeded with technical docs like Vue APIs, so you can import pre-vetted quizzes without uploading your own notes.

Key highlights

• Upload documents and select question types; the app generates questions from the text using an LLM (GPT-4 recommended).
• GPT evaluates answers against the original document, scoring correctness from 0 to 10.
• A bare-bones Ebbinghaus scheduler adjusts review dates based on those scores.
• Includes a question bank with pre-made quizzes for popular tech docs (e.g., Vue) that users can import.
• Runs locally via Docker, though it relies on external LLM APIs for generation and grading.

Caveats

• The author warns the project is in a very early stage and still contains many bugs.
• Updating is awkward: it requires manually exporting an .xlsx backup, rebuilding the Docker container, and re-importing the data.
• First-time page loads can be slow because of the module optimization scheme.

Verdict

Good for developers and students who want to stop hoarding notes and start retrieving them, provided you are comfortable self-hosting a Docker stack and feeding an LLM API. Skip it if you need a polished, zero-maintenance study tool.