DeepTutor: Why does an AI native learning system reach 20k stars in 7 months?

At the end of 2025, Chao Huang’s team from the Data Science Laboratory of Hong Kong University did something. They interviewed a large number of students and asked a simple question: What tools do you use to study now? The answers are scattered. Some people use Notion to take notes, some use Quizlet to answer questions, some open a session in ChatGPT as a tutor, and some feed textbooks into NotebookLM to generate summaries. There are many tools, but none of them can string together. Every time you change tools, the context is broken. After studying a chapter and want to take a self-test, you have to manually set the questions. After checking the information, I want to write a summary, so I have to switch to another editor. This kind of fragmentation is not an isolated phenomenon, but a common problem of current AI education tools. DeepTutor was forced out of these fragmented learning experiences.