Series: Deeptutor (1/7)
DeepTutor Beginner Course: 6 Articles to Build Your Own AI Learning Assistant
Have you ever run into this situation: you open ChatGPT, ask a question, get a great answer—but the next time you start a new conversation, it has no memory of what you said before. You have to re-introduce yourself, re-explain your background, and re-calibrate its understanding of you every single time.
Learning, above all else, requires someone who knows you. And almost every AI tool available today gets this terribly wrong.
DeepTutor was built to solve exactly this problem.
What Is DeepTutor
DeepTutor is an Agent-Native personalized learning assistant open-sourced by the Data Science Lab at the University of Hong Kong (HKUDS). "Agent-Native" means it's not just a chat box—it's a learning environment designed from the ground up around AI agents: it has memory, personality, proactively engages with you, and adapts its strategy based on your progress.
It has accumulated over 20,000 stars on GitHub, is licensed under Apache 2.0, and can be fully self-hosted, completely private, and entirely free.
What This Series Addresses
The six articles in this series are aimed at three types of people:
- Students and lifelong learners: who want an AI companion that truly understands their learning pace, not a general-purpose chatbot that needs to warm up every session
- Engineers and researchers: who need to feed large volumes of documents, papers, and technical references into an AI knowledge base for deep analysis and research
- Privacy-conscious users: who don't want to hand their learning data to cloud services and prefer to keep everything on their own machines
If you just want a quick sense of "is this worth my time," skip to the One-Line Summary at the end. If you want to fully understand the feature set and avoid common pitfalls, read all six articles in order.
Series Roadmap
1. What Is DeepTutor? Understanding the Design Philosophy of an Agent-Native Learning Assistant
Build the mental model first. Why traditional AI chat tools aren't suited for learning, what DeepTutor's design philosophy is, and where its fundamental differences from ChatGPT, Perplexity, and Notion AI lie.
2. DeepTutor Installation and Setup: From Zero to Running
Docker installation, native local installation, LLM backend configuration, and interface initialization—step by step until your environment is live. Includes common error handling.
3. DeepTutor's Five Modes Explained: Chat, Deep Solve, Quiz, Research, Math Animator
DeepTutor's five core learning modes, each designed for different use cases and strategies. This article breaks down all five—when to use which, and how to combine them for best results.
4. TutorBot: Build Your Personal AI Tutor with Memory, Personality, and Proactive Reminders
TutorBot is DeepTutor's most distinctive feature. This article covers how to create an AI tutor with its own teaching style, the design logic behind isolated memory, and how to connect it to Telegram, Discord, and other platforms so it reaches out to you.
5. Knowledge Base + RAG: Teaching DeepTutor Your Own Data
Upload PDFs, text files, and Markdown documents into DeepTutor's knowledge base, let it build semantic indexes with LightRAG, and call on them during conversations, research, and quizzes. This article covers knowledge base strategy and data organization best practices.
6. DeepTutor in the Real World: How Students, Engineers, and Researchers Actually Use It
The first five articles are the map; this one is the navigation. Three users from different backgrounds—each showing how to integrate DeepTutor into real learning and work workflows, with concrete templates and usage rhythms.
Recommended Reading Order
Follow the order 1 → 2 → 3 → 4 → 5 → 6: start with the design philosophy, get your environment set up, then go deeper into features, and finally map everything to real-world scenarios.
If time is limited, at minimum read 1 and 2: understand what it is, get it installed, and come back to the relevant sections when you're ready to explore specific features.
One-Line Summary
DeepTutor's core value is simple: it remembers you.
Most AI tools are just tools. Every time you use them, you have to re-teach them who you are, where your level is, and where you're headed. DeepTutor's design goal is to become your learning partner—from your very first conversation, it's accumulating an understanding of you, not resetting to zero every session.
If that matters to you, this series is worth reading seriously.