What DeepTutor is
DeepTutor is an open-source project that turns a large language model into a personalized tutor. Instead of one-shot answers, it keeps a model of the learner's level, gaps, and goals, and adjusts explanations and practice over time.
Why personalized tutoring is hard
Static courses move at one speed. A human tutor adapts in real time. DeepTutor tries to close that gap by tracking what a learner has mastered and what still needs work, then choosing the next step accordingly.
What it includes
- Adaptive explanations. Adjusts depth and examples to the learner's level.
- Practice and feedback. Generates questions and checks answers with hints.
- Progress tracking. Keeps a record of topics covered and gaps remaining.
- Self-hosted. Runs on your own infrastructure with your chosen model.
When it fits
DeepTutor suits self-directed learners and small teams who want a private, customizable tutor rather than a hosted SaaS. It is not a replacement for a qualified teacher in high-stakes settings.
Getting started
Follow the repository's setup guide, configure a model provider, and load a subject. Start with one topic to see how the adaptation behaves before expanding.
Repository: github.com/HKUDS/DeepTutor
