Skip to content

DeepTutor

What it is

DeepTutor is an AI-powered educational framework designed for personalized learning and intelligent tutoring, leveraging advanced reasoning models to guide students through complex topics.

What problem it solves

It addresses the limitations of standard chatbots in education by providing a structured, pedagogical approach to learning. It focuses on "teaching how to think" rather than just providing answers, using multi-turn reasoning to identify student misconceptions.

Where it fits in the stack

Category: Tool / AI Assistants & Knowledge

Typical use cases

  • Personalized coding tutor for students.
  • Intelligent assistant for complex technical documentation.
  • Research-driven educational agent experiments.

Strengths

  • Pedagogy-first design.
  • Leverages reasoning models for deep understanding of student intent.
  • Open-source framework for building educational agents.

Limitations

  • Requires high-reasoning models (e.g., GPT-4, Claude 3.5 Opus) for optimal performance.
  • Primarily a research/development framework rather than a finished consumer product.

When to use it

  • When building educational platforms that require more than simple Q&A.
  • When studying how agents can be used for persistent, structured learning.

When not to use it

  • When seeking a simple Q&A chatbot without the pedagogical overhead.
  • In production environments where latency and cost of multi-turn reasoning are primary concerns.

Getting started

Installation (Local)

DeepTutor requires Python 3.11+ and Node.js 20.9+.

  1. Clone & Setup: 
    git clone https://github.com/HKUDS/DeepTutor.git
cd DeepTutor
python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e ".[server]"
  2. Frontend: 
    cd web && npm install && cd ..
  3. Launch: 
    python scripts/start_web.py

Docker Deployment

docker compose -f docker-compose.ghcr.yml up -d

CLI Reference

DeepTutor provides a powerful agent-native CLI for interaction and management.

Interactive Chat

deeptutor chat --capability deep_solve --kb my-kb

One-shot Execution

deeptutor run deep_research "Transformer attention mechanisms"

Knowledge Base Management

deeptutor kb create my-kb --doc textbook.pdf
deeptutor kb add my-kb --docs-dir ./papers/

Architecture & Key Concepts

TutorBot

TutorBots are autonomous agents powered by nanobot. Each bot has its own workspace, memory, and "Soul" (personality template), allowing it to initiate proactive check-ins and evolve alongside the learner.

Book Engine

A multi-agent pipeline that transforms document collections into interactive "living books." It automatically generates outlines, retrieves relevant RAG context, and compiles pages with interactive block types (quizzes, animations, concept graphs).

Licensing and cost

  • Open Source: Yes (Apache-2.0)
  • API Usage: Dependent on the underlying LLM provider.

Sources / References

Contribution Metadata

  • Last reviewed: 2026-05-20
  • Confidence: high