github.com/GeminiLight/gen-mentor @main

<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/logo.png" alt="GenMentor Logo" width="300"/>

LLM-powered & Goal-oriented Tutoring System

<a href="https://www.tianfuwang.tech/gen-mentor">Website</a> &nbsp;·&nbsp;
<a href="https://arxiv.org/pdf/2501.15749">Paper</a> &nbsp;·&nbsp;
<a href="https://gen-mentor.streamlit.app/">Demo</a> &nbsp;·&nbsp;
<a href="https://youtu.be/vTdtGZop-Zc">Video</a>

---

[!IMPORTANT]
:sparkles: Welcome to visit the GenMentor website to learn more about our work!

This is official code of our paper "LLM-powered Multi-agent Framework for Goal-oriented Learning in Intelligent Tutoring System", accepted by WWW 2025 (Industry Track) as an Oral Presentation.

In this paper, we propose GenMentor, a large language model (LLM)-powered multi-agent framework designed for goal-oriented learning in Intelligent Tutoring Systems (ITS). This framework emphasizes personalization, adaptive learning, and goal-aligned content delivery, making it a robust solution for professional and lifelong learning scenarios.

🏫 ITS Paradigm Comparison

<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/its-paradigms.png" alt="GenMentor Overview" width="500" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px;"/>

Paradigm	Typical characteristics	Primary focus
🏫 Traditional MOOC	Static syllabus; pre-recorded lectures; fragmented learning	Broad access, low personalization
🤖 Chatbot ITS	Reactive Q&A; rule/LLM-driven; session-based help	Instant support, limited long-term adaptation
🎯 Goal-oriented ITS	Proactive planning; personalized paths; goal-aligned assessments	Targeted skill acquisition, continual adaptation

🤖 Key Agent Modules

<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor-framework.png" alt="GenMentor Overview" width="700" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px;"/>

• 🧭 Skill Gap Identifier: Analyzes learner's current knowledge to identify gaps.
• 👤 Adaptive Learner Modeler: Builds and updates learner profiles based on interactions.
• 🗓️ Learning Path Scheduler: Creates personalized learning paths and schedules.
• 📝 Tailored Content Generator: Produces customized learning materials and assessments.
• 🧑‍🏫 AI Chatbot Tutor: Engages learners in dialogue, answering questions and providing support.

Key Demo Interfaces

<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor_demo_1.png" alt="GenMentor Demo Interface-1" width="400" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px; margin: 8px;"/>
<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor_demo_2.png" alt="GenMentor Demo Interface-2" width="400" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px; margin: 8px;"/>
<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor_demo_3.png" alt="GenMentor Demo Interface-3" width="400" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px; margin: 8px;"/>
<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor_demo_4.png" alt="GenMentor Demo Interface-4" width="400" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px; margin: 8px;"/>
<img src="https://github.com/GeminiLight/gen-mentor/raw/main/resources/genmentor_demo_5.png" alt="GenMentor Demo Interface-5" width="400" style="box-shadow: 0 8px 24px rgba(0,0,0,0.15); border-radius: 8px; margin: 8px;"/>

🚀 Quick start

Here, we briefly describe how to set up and run the GenMentor system locally. Please see more details in the respective backend/ and frontend/ directories.

🛠️ Installation

1. Backend environment setup

cd backend
uv venv
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
uv pip install -r requirements.txt

1. Frontend environment setup

cd frontend
uv venv
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
uv pip install -r requirements.txt

⚙️ Configuration

1. Configure LLM secrets (at least one) for backend

Option A: update a .env file in backend/

# Example for OpenAI:
OPENAI_API_KEY="your-openai-api-key"

# Example for DeepSeek:
DEEPSEEK_API_KEY="your-deepseek-api-key"

Option B: export environment variables in your shell

# Example for OpenAI:
export OPENAI_API_KEY="your-openai-api-key"

# Example for DeepSeek:
export DEEPSEEK_API_KEY="your-deepseek-api-key"
DEEPSEEK_API_KEY="your-deepseek-api-key"

1. Configure api endpoint for frontend

If you would like to run the backend on a different host/port, please update the API URL in frontend/config.py:

backend_endpoint = "http://127.0.0.1:5000/"

▶️ Running Locally

[!NOTE] The default ports are 5000 for backend, 8501 for frontend by default

Option A: Manual (preferred when using separate venvs)

# start backend
cd backend
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
uvicorn main:app --reload --port 5000

# start frontend
cd frontend
source .venv/bin/activate  # on Windows: .venv\Scripts\activate
streamlit run main.py --server.port 8501

Option B: Helper scripts (single shell; assumes uvicorn/streamlit on PATH)

# start backend
bash ./scripts/start_backend.sh [PORT]

# start frontend
bash ./scripts/start_frontend.sh [PORT]

# stop all
bash ./scripts/stop_all.sh

🌐 Accessing the App

Finally, you can access:

• Backend API: http://127.0.0.1:5000/
• Frontend UI: http://127.0.0.1:8501/

🚀 Demo Version of Web Application

Welcome to explore the demo version of the GenMentor web application:

GenMentor Web App

This interactive demo showcases GenMentor's core functionalities, including:

• Skill Gap Identification: Precisely map learner goals to required skills.
• Adaptive Learner Modeling: Capture learner progress and preferences.
• Personalized Content Delivery: Generate tailored learning resources.

You could also watch the demo video for a quick overview (click the image below):

📚 Citation

@inproceedings{wang2025llm,
  title={LLM-powered Multi-agent Framework for Goal-oriented Learning in Intelligent Tutoring System},
  author={Wang, Tianfu and Zhan, Yi and Lian, Jianxun and Hu, Zhengyu and Yuan, Nicholas Jing and Zhang, Qi and Xie, Xing and Xiong, Hui},
  booktitle={Companion Proceedings of the ACM Web Conference},
  year={2025}
}