Personalized Learning Mentor: MentorAI 🎓

A multi-agent AI system that builds bespoke 30-day curricula, adapts to your learning style, and ensures mastery through real-time multimodal coaching.

---

1. The Problem 🚨

Self-directed learning is often overwhelming and inefficient due to:

• Too much content with no clear path (“analysis paralysis”).
• Lack of personalized guidance and feedback.
• Static curricula that fail to adapt to prior knowledge and learning style.

2. The Solution 💡

MentorAI is a multi-agent AI system that acts as a personalized tutor. It:

• Builds a 30-day adaptive curriculum for each learner.
• Maintains long-term memory of skills, goals, and learning style.
• Interacts via text, voice, and visuals, providing guidance, quizzes, and feedback.

MentorAI empowers learners by providing structured, adaptive learning plans that improve engagement and skill mastery without the need for human tutors.

---

3. Value Proposition

• For Learners: Provides a clear, structured, and adaptive path to mastering new skills, moving beyond static content to offer a truly personalized tutoring experience.
• For Educators: Offers a scalable model for delivering one-on-one coaching, ensuring every student gets the support they need to succeed.
• For Developers: Serves as a comprehensive, open-source example of building a complex, stateful multi-agent system with Google's Gemini models.

---

4. Core Features

• 🤖 Multi-Agent System: Orchestrates five specialized AI agents, each powered by a specific Gemini model, to create a seamless learning journey.
• 🧠 Adaptive Curriculum: Generates a personalized 30-day learning plan based on an initial diagnostic interview, not just self-reported skill levels.
• 💬 Multimodal Coaching: Interacts via text, voice, and vision. Users can upload screenshots of code, ask questions verbally, and receive visual explanations like diagrams.
• 📝 Daily Quizzes & Feedback: Reinforces learning with daily quizzes and provides targeted feedback to address misconceptions.
• 💾 Long-Term Memory: Stores user progress, goals, and learning preferences locally to maintain context across sessions.

---

5. Architecture & Agent Design

The application orchestrates 5 Specialized Agents to guide the user from "Zero to Mastery".

---

This project leverages the specific strengths of the Gemini model family via the @google/genai SDK.

🧠 1. The Diagnostic Agent ("The Gatekeeper")

• Model: gemini-1.5-flash
• Role: Adaptive Interviewer.
• Innovation: Instead of relying on self-assessment (e.g., "I'm intermediate"), it uses "Gatekeeper Scenarios"—concrete problems to test actual competence using Bloom's Taxonomy.

🗺️ 2. The Planner Agent ("The Architect")

• Model: gemini-1.5-pro-latest
• Role: Curriculum Designer.
• Innovation: Uses a large context window and advanced reasoning to deeply analyze the user's profile, ensuring the 30-day plan is coherent and logically sequenced. It outputs a structured JSON object for the curriculum.

🎓 3. The Coaching Agent ("The Companion")

• Model: gemini-1.5-pro-latest
• Role: Socratic Tutor.
• Innovation: Fully Multimodal.
  • Vision: Can analyze uploaded screenshots of code errors or handwritten notes.
  • Visuals: Generates diagrams using a Hybrid Renderer that attempts Mermaid.js first and falls back to React Flow for syntax errors.
  • Grounding: Connects to Google Search for up-to-date documentation.

🎙️ 4. The Voice Agent ("The Conversationalist")

• Model: gemini-1.5-pro-latest
• Role: Verbal Tutor.
• Innovation: Connects directly to the Gemini API via WebSockets. It handles raw PCM audio encoding/decoding in the browser for low-latency, interruptible voice conversations.

📊 5. The Feedback Loop

• Models: gemini-1.5-flash
• Role: Assessor.
• Innovation: The Quiz Agent generates questions based strictly on that day's objective. The Feedback Agent analyzes mistakes to identify root misconceptions and provides direct links to remedial resources.

---

6. Key Concepts Implemented

This project demonstrates several key concepts from the "Build with Agents" course:

1. 🤖 Agentic Mindset & Task Decomposition: The problem of personalized learning is broken down into specialized tasks, each handled by a dedicated agent (Diagnostic, Planner, Coach, etc.).
2. ⛓️ Chaining & Orchestration: The agents work in a logical sequence. The output of the Diagnostic Agent (Learner Profile) is chained as the input for the Planner Agent, which then creates the curriculum used by the Coaching Agent.
3. 🛠️ Function Calling / Tools: The Coaching Agent is equipped with tools like Google Search (for grounding) and a Mermaid Diagram Generator (for visual explanations), enhancing its capabilities beyond simple text generation.
4. 🔮 Multimodality: The Coaching and Voice agents showcase multimodal interactions, accepting user input via text, images (vision), and audio, and providing visual output via diagrams.

---

7. Installation & Setup

Prerequisites

• Node.js 18+ installed.
• A Google AI Studio API Key.

Local Development

1. Clone the repository

   git clone https://github.com/your-username/MentorAI.git
   cd MentorAI

2. Install Dependencies

   npm install

3. Configure API Key Create a .env.local file in the root of the project and add your Google AI Studio API key:

   # .env.local
   API_KEY=your_actual_api_key_here

   Alternatively, you can set the environment variable directly in your terminal (PowerShell on Windows):

   $env:API_KEY="your_actual_api_key_here"
   npm run dev

4. Run the App

   npm run dev

   Open http://localhost:5173 in your browser. (or the port shown in your terminal)

Docker Deployment (Optional)🐳

To build and run the application in a Docker container:

1. Build the Image

   docker build -t mentor-ai .

2. Run the Container Pass your API key as an environment variable and map the port:

   docker run -p 8080:5173 -e API_KEY="your_key_here" mentor-ai

   Access the application at http://localhost:8080.

---

8. Conclusion

MentorAI represents a significant step forward in personalized education. By leveraging a sophisticated multi-agent architecture, it directly addresses the common pitfalls of self-learning—information overload, lack of guidance, and static curricula.

It's more than just a learning tool; it's a blueprint for creating truly adaptive, engaging, and effective educational experiences powered by generative AI. This project not only empowers individual learners but also demonstrates the immense potential of agentic AI to revolutionize how we acquire and master new skills.

---

9. Future Roadmap

• Real-time Code Execution: Integrating Pyodide to run Python code directly in the browser instead of just simulating the output.
• RAG Integration: Adding a Vector Database (e.g., Pinecone) to store long-term conversation history, allowing the agent to reference mistakes made weeks ago.
• LMS Integration: Adding LTI support to plug into Canvas/Blackboard for institutional use.

---