🧠 MentorAI — Personalized, Agentic Education for Every Learner

MentorAI is an intelligent, multilingual teaching agent that personalizes education for each student — from entrance exam to adaptive lessons, tests, and feedback — all delivered in the student's preferred language and learning style. Designed to empower under-resourced learners in rural and diverse linguistic communities.

---

🚀 Demo

🎥 Watch Demo Video

🌐 Live App 📸

---

💡 Problem

Millions of students, especially in rural or underserved communities, face:

• 🚫 Lack of teachers or personalized attention
• 🌐 Language barriers in English-centric education
• 🧹 Varying learning speeds and no adaptive pacing
• 😟 Demotivation due to unfair comparisons with others

---

✅ Solution

Mentora uses LLM Agents + Voice + Language Adaptation to:

• 🧪 Assess student’s current level through an entrance exam
• 📺 Create a personalized learning path based on results
• 🎓 Deliver lessons in voice + text, adapted to the student’s preferred language
• 📝 Generate interactive quizzes to assess understanding
• ♻️ Continuously adapt the curriculum based on student performance
• 📊 Track progress via charts for teachers
• 🔐 Support user authentication for teachers/students
• 🤝 Encourage progress while removing pressure to compare

---

🧠 How It Works

graph TD
    A[Student Begins] --> B(Entrance Assessment Agent)
    B --> C(Curriculum Agent: Builds Lesson Plan)
    C --> D(Teaching Agent: Voice + Text Explanation)
    D --> E(Quiz Agent: Assess Learning)
    E --> F(Feedback Agent: Personalized Review)
    F --> G(Adjusts Curriculum to Student Needs)
    G --> C(Engagement Agent: Encourages Individual Student Progress)

Loading

---

⚙️ Tech Stack

Component	Tool
💬 LLM Engine	Anthropic Claude via API
⚡ Inference	Groq API for ultra-fast LLM interaction
🧠 Agent Orchestration	Fetch.ai Agentverse
♻️ Workflow Coordination	Orkes Conductor
😤 Voice Output	LettA Pro / Google TTS
🧯 Speech Input	Whisper / Google Speech API
💻 Frontend	Streamlit / React
🧪 Mock Backend	Vappi API mocks for student data
📦 LMS Data Storage	Vappi API mocks for student + teacher records
📊 Visualization	Plotly / Matplotlib
🔐 Authentication	Firebase Auth / Custom Flask backend with bcrypt

---

🌍 Use Case Scenario

1. Student speaks Hindi and struggles with English grammar.
2. MentorAI assesses their level with a few voice-based questions.
3. It builds a lesson plan starting with “Simple Past Tense.”
4. Lesson is delivered in Hindi with voice + text.
5. Quiz evaluates understanding.
6. Feedback loop adjusts the curriculum and offers encouragement.
7. Teacher views progress in a chart dashboard.

---

📁 Folder Structure

MENTORAI/
├── agents/
│   ├── assessment_agent.py
│   ├── curriculum_agent.py
│   ├── teaching_agent.py
│   ├── feedback_agent.py
│   ├── quiz_agent.py
├── frontend/
│   └── app.py (Streamlit UI)
│   └── dashboard.py (Teacher Dashboard)
├── voice/
│   ├── tts.py
│   ├── stt.py
├── utils/
│   └── language_utils.py
│   └── student_data.py
│   └── auth.py
├── api/
│   ├── vappi_client.py (mock LMS integration)
├── assets/
│   └── demo-video.mp4
└── README.md

---

🔧 Running Locally

1. Clone the repo:

   git clone https://github.com/ABCoder1/mentorAI.git
   cd mentorAI

2. Install dependencies:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`
   pip install -r requirements.txt

3. Create a .env file and add your API keys:

   CLAUDE_API_KEY=...
   GROQ_API_KEY=...
   GOOGLE_TTS_KEY=...
   VAPPI_TOKEN=...
   FIREBASE_KEY=...
   

4. Run the app:

   streamlit run frontend/app.py

---

💬 Sample Prompt Flow (Claude API)

{
  "input": "Please assess this student's understanding of basic grammar. Here are their answers: ...",
  "instruction": "Return a level, suggested next topic, and 3 quiz questions."
}

---

📈 Teacher Dashboard

🔐 Teacher logs in to their account

🧑‍🏫 Registers course content and language options

📊 Views line/bar charts of student progress using Plotly

📁 Downloads report cards (CSV/PDF)

---

🔐 Authentication Features

Student/Teacher registration via email/password

Role-based routing (Teacher dashboard / Student dashboard)

Secure session using Firebase or Flask auth

---

📦 Vappi LMS Mock Usage

POST /student/record – Save performance

GET /student/{id} – Fetch profile + score history

POST /course/register – Teacher creates new course

---

📈 Future Vision

• 🧠 Add memory to track long-term progress
• 🧪 Train with open datasets to teach STEM, coding (DP, Graphs, etc.)
• 🌐 Expand to 10+ Indian languages

---

🙌 Team & Acknowledgements

• Aditya Bhardwaj — Architect
• Anant Patel — Developer
• Pankaj Sharma — Vision
• Thanks to UC Berkeley, Fetch.ai, Anthropic, Groq, Orkes, and Vappi for tools & support.

---

🏆 Submission Track

🧠 Social Impact + 🛠️ Productivity in Education

---

📢 Contact

Email: [abhar061@ucr.edu] GitHub: [github.com/ABCoder1]