About LearnMate
LearnMate is an MCP server in the RAG category: this system functions as a domain-specific autonomous AI agent, using agent orchestration frameworks capable of contextual reasoning across multiple books, multimodal inputs and memory driven personalized learning.
Install
git clone https://github.com/priyanshi0609/LearnMateThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Ask AI
Ask AI about LearnMate
Powered by Claude · Grounded in docs
Security checks
- ✓README presentThe repository ships documentation.
5 checks have not been run yet.
README
🚀 LearnMate – AI-Powered Personalized Learning Copilot
A smart, AI-driven platform that creates personalized revision plans, identifies weak areas from test data, and assists students with doubt resolution using Google Gemini, LangChain, and Vector Search.
📌 Table of Contents
- Overview
- Features
- Tech Stack
- Architecture
- Demo
- How It Works
- Installation
- Usage
- Future Scope
- Contributors
- License
🧠 Overview
LearnMate is a personalized AI learning assistant designed to revolutionize how students revise and master subjects. It uses test performance analysis, GPT-4 reasoning, and Retrieval-Augmented Generation (RAG) to generate smart revision plans and answer topic-based queries using a semantic search system.
✨ Features
✅ Upload your test results (CSV or manual entry)
✅ AI detects weak topics and generates a 7-day revision plan
✅ Ask doubts and get accurate explanations from your own notes
✅ Semantic search powered by embeddings + Vector DB
✅ User-friendly UI for dashboard and chat
✅ AI memory to track what you've studied (WIP)
🛠️ Tech Stack
| Layer | Technology |
|---|---|
| AI/LLM | Google Gemini, LangChain |
| Vector Search | ChromaDB / Pinecone |
| Backend Logic | Python, LangChain Agents & Tools |
| File Handling | Pandas, PyMuPDF (for PDF parsing) |
| UI | React and Tailwind CSS |
| Embeddings | OpenAI / HuggingFace Sentence Transformers |
🏗️ Architecture
Student Input (CSV/Text)
↓
Gemini via LangChain (Analysis & Plan Generation)
↓
Personalized Study Plan
↓
Ask Doubts ➝ Query Embeddings ➝ ChromaDB ➝ Relevant Context ➝ Gemini Answer
⚙️ How It Works
1. Upload Performance Data
- Accepts CSV files with “Subject” and “Score” columns
- Filters weak topics (e.g., Score < 60)
2. GPT-4 Study Plan Generator
- Prompts GPT to create a customized 7-day plan
- Plans include tasks: reading, solving, revising
3. Ask Doubts (RAG Pipeline)
- Upload your notes as PDFs or text files
- Chunked, embedded, and stored in ChromaDB
- GPT retrieves relevant content → answers your queries
🚀 Installation
git clone https://github.com/priyanshi0609/LearnMate.git
cd learnmate
pip install -r requirements.txt
Ensure your .env contains:
GOOGLE_API_KEY=your-key-here
To run (Backend)
cd backend
$env:PYTHONPATH="."; uvicorn app:app --reload
💡 Usage
- Upload a test result file or enter topics manually
- Click “Generate Plan” to view the AI-created revision plan
- Use the “Ask Doubts” feature to ask academic questions
- View your responses and daily tasks in a clean dashboard
🔮 Future Scope
- 🧠 Add LangChain memory to track user progress
- 📘 Weekly summary + flashcard generation
- 📊 Integration with Google Classroom or MS Teams
- 🗂️ AI-generated quizzes with answer explanations
👩💻 Contributors
- Priyanshi Bothra – @priyanshibothra
📄 License
This project is licensed under the MIT License – see the LICENSE file for details.
README copy captured when this server was indexed; it may lag the repository. The original is authoritative.