Meducate MCP

Meducate MCP

A self-aware, AI-powered MCP server that acts as a personal medical education assistant for African students, understanding natural language and autonomously managing reminders, data bundles, and medical knowledge retrieval.

Category
访问服务器

README

🏥 Meducate MCP - Africa's AI Medical Education Assistant

"Your AI-powered Jarvis(let's call him Ananse) for medical education in Africa"

Empowering African medical students and young doctors with intelligent, context-aware assistance through natural language AI and Model Context Protocol.

Link to Video Demo

Link to Presentation Slides


🌍 The African Medical Education Challenge

Meet Serwaa, a second-year medical student in Accra, Ghana:

  • 📚 Overwhelmed with lectures, labs, and clinical rotations
  • 💸 Limited budget for textbooks and study resources
  • 📱 Struggles to afford consistent internet for research
  • ⏰ Needs constant reminders but finds rigid apps frustrating
  • 🌐 Often searches for medical information in broken English or local languages

Serwaa's reality mirrors millions of medical students across Africa - brilliant minds facing systemic barriers that technology should solve, not complicate.


✨ Enter Meducate MCP

Meducate is a self-aware, AI-powered Model Context Protocol (MCP) server that acts like a personal medical education assistant - understanding natural language, working autonomously in the background, and bridging critical gaps in African healthcare education.

🎯 What Makes Meducate Revolutionary

  1. 🧠 Self-Aware & Autonomous

    • Constantly monitors your schedule and study needs
    • Understands natural language: "remind me every 5 minutes" or "quiz me tomorrow morning"
    • Works in the background - no constant manual input needed
  2. 🌍 African-First Design

    • First MCP integration with African telecoms (MTN, Airteltigo)
    • Buy data bundles at cheap prices directly through AI in natural language.
    • Track balances, send usage stats to guardians
    • First Hubnets MCP for SMS/OTP in Africa
  3. 📖 Medical Knowledge at Your Fingertips

    • Scrapes trusted sources: WHO Africa, CDC, PubMed, BMJ
    • Focuses on diseases relevant to African contexts (Malaria, TB, etc.)
    • Generates quizzes tailored to your curriculum
    • Finds cheapest textbook prices across platforms
  4. 🗣️ Natural Language Understanding

    • Write in English, Twi, Pidgin, or local languages
    • AI interprets context, not just keywords
    • Accessible to non-native speakers

🚀 Key Features

🤖 AI-Powered Intelligence

  • Natural Language Processing via Groq AI (Llama 3.3 70B)
  • Dynamic tool selection - AI decides which action to take
  • Context-aware suggestions based on your study patterns
  • Tools are automatically discovered and integrated into the AI's capabilities - no hardcoding required!

📅 Smart Scheduling System

"Quiz me about pharmacology tomorrow morning"
"Remind me every hour to drink water"
"Send my study stats to mom every Sunday"
"buy 5gb for my data bundle when my balance is low"
"send 1gb to my sister tomorrow afternoon to thank her for helping me study`"

The AI understands and executes - no rigid formats!

📚 Medical Knowledge Base

  • Disease information scraping (WHO, CDC, PubMed)

📚 Medical Knowledge Base

  • Disease information scraping (WHO, CDC, PubMed)
  • Quiz generation for active recall
  • Study resource aggregation (videos, articles)
  • Textbook price comparison

📱 Africa-Specific Integrations

  • Data bundle purchases (Airteltigo, MTN)
  • SMS notifications via Hubnets
  • Gmail notifications for reminders and updates

📊 Productivity Tools

  • Google Sheets integration for timetables
  • Gmail notifications
  • Autonomous/self-conscious reminder system

🏆 Why Meducate Wins

✅ Technical Excellence

  • JSON-based MCP protocol implementation
  • Multi-service architecture (Python + Node.js Puppeteer)
  • AI-powered natural language parsing
  • Dynamic tool execution system
  • Robust error handling

✅ Innovation

  • First African telcom MCP integration
  • First Hubnets SMS MCP
  • Natural language scheduler (vs rigid cron jobs)
  • Self-aware autonomous operation
  • Multi-domain integration (healthcare + telecom + education)

✅ African Relevance

  • Solves real problems facing African medical students
  • Focus on African disease contexts
  • Affordable data management built-in
  • Accessible natural language interface

🎬 Quick Start

Prerequisites

  • Python 3.10+
  • Node.js 18+
  • Google Cloud account (for Sheets/Gmail)
  • Groq API key(This is free, you can get it via https://console.groq.com/keys)

Installation

  1. Clone the repository
git clone https://github.com/JoshNuku/meducate-mcp.git
cd meducate-mcp
  1. Set up Python environment
# Install UV package manager
pip install uv

# Install dependencies
uv sync
  1. Set up Puppeteer server
cd puppeteer-server
npm install
  1. Configure environment
# Copy example environment file
cp .env.example .env

# Edit .env with your credentials
# Add GROQ_API_KEY, Gmail credentials, etc.
  1. Set up Google Sheets API

    • Create a project in Google Cloud Console
    • Enable Google Sheets API
    • Create service account and download credentials.json
    • Place credentials.json in the root directory
  2. Run the servers

# Terminal 1: Start Puppeteer server
cd puppeteer-server
npm start

# Terminal 2: Start MCP server
cd ..
python main.py

📖 Usage Examples

Natural Language Reminders

"Remind me to review anatomy every morning at 8am"
"Quiz me about cardiology tomorrow afternoon"
"Send my study hours to dad every Sunday"
"Drink water every 30 minutes"

Medical Information

"Get latest WHO guidelines on malaria treatment"
"Find recent research on tuberculosis in Africa"
"What are common symptoms of typhoid fever?"

Data Management

"Buy 5GB data bundle for my MTN number"
"Check my data balance"

Study Resources

"Find YouTube videos about pharmacology"
"Compare prices for Gray's Anatomy textbook"
"Generate 10 quiz questions on neurology"

🌍 African Language Support

Meducate understands context in multiple languages:

English (Standard)

"Remind me to study at 6pm everyday"

Pidgin (Nigeria)

"Make you remind me say make I read book for evening time"

Twi (Ghana)

"kai me s3 mefa "(Remind)

Swahili (East Africa)

"Nikumbushe kusoma saa kumi na mbili usiku"

The AI interprets intent, not just exact phrasing!


🏗️ Architecture

┌─────────────────────────────────────────────────┐
│          User (AI Assistant/LLM)                │
└────────────────┬────────────────────────────────┘
                 │ MCP Protocol (JSON-RPC)
                 ▼
┌─────────────────────────────────────────────────┐
│         Meducate MCP Server (Python)            │
│  ┌──────────────────────────────────────────┐  │
│  │  AI Decision Engine (Groq/Llama 3.3)     │  │
│  │  - Natural language parsing              │  │
│  │  - Dynamic tool selection                │  │
│  │  - Context-aware reasoning               │  │
│  └──────────────────────────────────────────┘  │
│  ┌──────────────────────────────────────────┐  │
│  │  Tools Layer                             │  │
│  │  • Medical scraping  • Reminders         │  │
│  │  • Telecom APIs      • Payments          │  │
│  │  • Notifications     • Scheduling        │  │
│  └──────────────────────────────────────────┘  │
└──┬──────────────┬──────────────┬───────────────┘
   │              │              │
   ▼              ▼              ▼
┌──────────┐  ┌──────────┐  ┌──────────────┐
│Puppeteer │  │  Google  │  │   African    │
│ Server   │  │ Services │  │   Telecom    │
│(Node.js) │  │(Sheets,  │  │    APIs      │
│          │  │ Gmail)   │  │ (MTN, etc.)  │
└──────────┘  └──────────┘  └──────────────┘
     │
     ▼
┌─────────────────────────────────────┐
│  Medical Websites                   │
│  • WHO Africa  • CDC  • PubMed      │
│  • BMJ  • UpToDate  • Book Stores   │
└─────────────────────────────────────┘

📁 Project Structure

meducate_mcp/
├── main.py                      # Entry point & MCP server
├── .env.example                 # Environment template
├── credentials.json             # Google API credentials
├── pyproject.toml               # Python dependencies
│
├── Tools/                       # MCP tool implementations
│   ├── disease_tools.py         # Medical info scraping
│   ├── quiz_tools.py            # Quiz generation
│   ├── study_tools.py           # Study resources
│   ├── notification_tools.py    # Email/SMS alerts
│   ├── reminder.py              # Reminder system
│   └── buy_internet_bundle.py   # Data bundle purchases
│
├── utils/                       # Core utilities
│   ├── ai_decision_engine.py    # Groq AI integration
│   ├── scheduler.py             # Background scheduler
│   ├── tool_extractor.py        # Dynamic tool discovery
│   └── message_formatter.py     # AI message formatting
│
├── Helpers/                     # Business logic(functions for tools)
│   ├── scheduler_helper.py      # Scheduling functions
│   ├── notification_helpers.py  # Notification logic
│   └── reminder_helpers.py      # Reminder helpers
│
├── puppeteer-server/            # Web scraping service
│   ├── src/
│   │   ├── controllers/         # Scraping controllers
│   │   ├── services/            # Browser service
│   │   └── utils/               # Page utilities
│   └── package.json
│
└── at root/                        # Documentation
    ├── NOTIFICATION_DOCS.md
    ├── PUPPETEER_TOOLS_README.md
    └── REMINDER_DOCS.md

🎯 MCP Hackathon Africa 2025

Tracks Covered

  • 🏥 Healthcare Systems (Primary)
  • 📚 Education & Learning (Primary)
  • 📡 Telecommunications (Secondary)

Unique Value Propositions

  1. Only MCP with African telecom integration

    • Direct data bundle purchases- no stress of dialing ussd codes
    • Balance tracking and alerts
    • Guardian notifications
    • email sending- the course rep who has to announce to the class about rescheduled lectures can do so via gmail notification tool integrated into the mcp
  2. First Unofficial MCP with Hubnets SMS

    • OTP verification
    • SMS reminders for no-internet scenarios
    • Wider accessibility
  3. Natural Language AI Scheduler

    • Not just keyword matching
    • True context understanding
    • Works in multiple languages
  4. Self-Aware Operation

    • Autonomous background monitoring
    • Proactive suggestions
    • Contextual awareness
  5. Medical Education Focus

    • African disease contexts
    • Affordable resource discovery
    • Quiz generation for active learning

📚 Documentation


🌟 Impact & Vision

Current Impact

Reduces study costs by aggregating free medical resources
Saves data expenses through affordable bundle integration
Increases efficiency with AI-powered autonomous reminders
Improves accessibility via natural language understanding
Bridges telecom gap with first-of-its-kind African MCP integrations

Future Vision

🚀 Expand to more African countries and telecom providers
🎙️ Add voice interface for hands-free, low-literacy operation
🏥 Integrate with hospital systems for clinical rotation management
📚 Build community knowledge base of African medical cases
🤝 Partner with medical schools across Africa
🌍 Support more local languages (Yoruba, Igbo, Amharic, etc.)
💡 AI-powered study group matching and collaboration


🤝 Contributing

We welcome contributions! This is an open-source project for African impact.

See CONTRIBUTING.md for detailed developer guidelines.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选