HKUDS Ships DeepTutor, Passes 196 GitHub Stars

HKUDS shipped DeepTutor and it's already past 196 GitHub stars, per the project's repo. That makes it the lab's third agent project to gain real traction, not its first.

DeepTutor targets a different problem than most tutoring bots. Rather than reset context every session, it holds a persistent model of a single learner over time, running on HKUDS's own retrieval stack. The idea is a study partner that remembers where you left off, not a chatbot you re-explain yourself to every time you open it.

The repo doesn't say how DeepTutor tracks what a learner already knows or how that model updates as someone improves. Anyone deciding whether to build on it will need to dig past the star count for that.

What matters isn't 196 stars by itself. That's a solid showing, not a viral one. It's that HKUDS keeps shipping agent projects people actually adopt, without a press team behind every release. Three hits from one academic group is a better signal for what to build on than a big lab's next launch.