Skip to main content
DeepTutor is an open-source AI learning platform that turns your documents into an intelligent tutor. Upload textbooks, research papers, or technical manuals and immediately start solving problems, generating practice exams, conducting deep research, and creating interactive learning visualizations — all powered by a multi-agent system with RAG, knowledge graphs, and real-time web search.

Quick start

Get DeepTutor running locally in minutes with Docker or manual install.

Configuration

Configure your LLM provider, embedding model, and search tools.

Core features

Explore the knowledge base, smart solver, research, and more.

API reference

Integrate DeepTutor’s backend into your own applications.

​
What you can do with DeepTutor

Smart solver

Ask questions about your documents and get step-by-step answers with exact citations from a dual-loop multi-agent reasoning system.

Question generator

Auto-generate quizzes and practice problems from your knowledge base, or clone the style of a reference exam.

Deep research

Conduct systematic literature reviews across your documents, the web, and academic paper databases with automatic citation management.

Guided learning

Generate interactive visual explanations of complex concepts from your notebook, with context-aware Q&A at every step.

Idea generation

Discover novel research directions and synthesize cross-domain insights from your learning materials.

Co-writer

AI-assisted Markdown editor with rewrite, expand, and narration features powered by RAG and web search.

​
Get up and running

1

Install DeepTutor

Clone the repository and copy the environment template:
git clone https://github.com/HKUDS/DeepTutor.git
cd DeepTutor
cp .env.example .env
2

Configure your LLM

Edit .env with your API key and model settings. At minimum, set LLM_API_KEY, LLM_HOST, LLM_MODEL, EMBEDDING_API_KEY, EMBEDDING_HOST, and EMBEDDING_MODEL.
3

Launch with Docker

docker compose up
The frontend is available at http://localhost:3782 and the API at http://localhost:8001.
4

Create your first knowledge base

Go to http://localhost:3782/knowledge, click New Knowledge Base, and upload your PDF, TXT, or Markdown files to start learning.
DeepTutor supports both Docker and manual Python/Node.js installation. See the deployment guides for full options including cloud and custom port configurations.

Build docs developers (and LLMs) love

Get started for freeTalk to us