DeepTutor
ActiveDescription
DeepTutor: Agent-Native Personalized Learning Assistant.
Key Features
- Three-layer memory (L1/L2/L3) — Short-term session, mid-term working, long-term semantic memory with cross-session persistence
- TutorBot persistent autonomous tutor — Long-running autonomous AI teacher with HTTP/SSE API and per-user isolation
- RAG retrieval augmentation — LlamaIndex-based document indexing and retrieval with versioned KB indexes and auto-reindex
- Multimodal interaction — Text, code, SVG attachments with Visualize/Animator visualization engine integration
- Agent-native architecture — Agentic capabilities like Deep Research / Solve / Question with Auto Mode for tool selection
- Multi-user deployment — Admin grants, per-user workspace isolation and scoped runtime access control
Use Cases
💡 University course personalized tutoring with RAG retrieval from course materials for deep Q&A
💡 Enterprise training platform with TutorBot as a long-running AI tutor providing continuous learning support
💡 Programming learning assistant supporting code attachment input, execution feedback and interactive debugging
💡 K-12 education scenarios converting abstract concepts into interactive diagrams via visualization engine
💡 Multi-user collaborative learning environments with independent workspaces and memory per user
Strengths & Limitations
✅ Strengths
- • Actively maintained, recent updates
- • High community interest (35.9k stars)
- • Permissive open-source license (Apache-2.0)
- • Responsive to issues, low backlog
Categories
Quick Start
```bash
git clone https://github.com/HKUDS/DeepTutor.git
cd DeepTutor
pip install -r requirements.txt
# Configure LLM API Key in .env
deptutor start
```