Inspiration

In Sudan, many lives have been lost or Endangered simply due to a lack of first aid awareness. During times of conflict and daily emergencies alike, people often act with good intentions but without the right knowledge - offering food to someone who’s been starving, applying harmful home remedies to burns, or failing to manage bleeding or shock. These small mistakes, made in critical moments, can lead to devastating outcomes.

Our inspiration to create Mongez came from real stories we witnessed and heard, people who tried to help, but didn’t know how. We wanted to change that, we believed that if we could give every Sudanese person access to simple, lifesaving information, we could make a real difference.

What it does

Mongez is a user-friendly first aid assistant designed specifically for Sudanese communities. It provides step-by-step guidance for handling medical emergencies based on the type of injury or condition, the severity, and the user’s location.

Mongez features:

Chatbot Guidance: available in both English and Sudanese Arabic, the Chatbot interacts naturally with users, offering easy-to-follow first aid instructions.
Customized Recommendations: Based on user location, Mongez suggests the nearest appropriate medical facility from a database of hospitals and clinics across Sudan.
Doctor Volunteer page: the platform includes a section for doctors and volunteers who wish to offer remote assistance, building a support network.
Culturally Sensitive Responses: All responses are localized, and the Arabic chatbot speaks in a Sudanese dialect, making it easier for users to understand and trust.
Easy Access: The services are currently available through a webpage and also via telegram, offering wide accessibility.

How We Built Monqez

AI-Powered with RAG Architecture: At the heart of Monqez is a Retrieval-Augmented Generation (RAG) system. We combined Google Gemini with LangChain and FAISS to enable the assistant to retrieve real, verified Sudanese first aid guidance from PDF documents and generate medically informed, context-aware responses. This hybrid approach ensures answers are both accurate and up-to-date.

Location-Aware Emergency Guidance: Monqez uses the user’s real-time GPS coordinates to identify and recommend the 7 closest hospitals from a curated GeoJSON dataset of medical facilities in Sudan. This ensures faster decision-making in critical moments.

Real-Time Volunteer Matching: We built a live system that connects users to registered volunteer doctors, filtered by specialty, availability, and timing. This makes professional help accessible even when formal healthcare systems are overwhelmed or out of reach.

Multichannel Chat Access: Beyond the web app, Monqez integrates with Telegram, enabling users in remote or low-connectivity areas to interact with the bot and get emergency help with just a few taps.

Backend Stack: Built on Flask, a lightweight Python web framework, with a modular, RESTful structure. User sessions, hospital data, and chat histories are handled using SQLAlchemy for ORM and data persistence.

Frontend and User Interface: A responsive interface built using HTML, CSS, and Bootstrap, optimized for both mobile and desktop. We focused on clarity, speed, and minimal distractions to avoid user confusion in emergencies.

Tools and Libraries

LangChain and FAISS for vector search and document retrieval
HuggingFace Transformers and PyTorch for embedding generation
Google Gemini API for contextual AI responses
Flask-Login, Flask-WTF, and Flask-Session for user management and security
python-dotenv and Render for deployment and environment handling

Challenges we faced

poor internet access
lack of local data
Time pressure

What next for Mongez

Emergency Button : a quick button to send your location and call for help.
Voice Help : Spoken instructions in Sudanese Arabic, will be useful for the blind and elderly.
Works Without Internet : can use saved guides when offline.
Blood Donation Feature : connects people who need blood with donors nearby.
Hospital Directions : shows fastest way to hospitals.