Inspiration
AtlasRoute was inspired by a deeply personal real-life experience that happened while I was still learning Electronic Engineering in secondary school and working part-time after school hours on electrical installation projects.
We were assigned to a housing installation project in Rivers State, Port Harcourt, Nigeria. The project required us to stay in a rural environment for about three weeks, working on-site daily. After work, we often relied on nearby houses for electricity just to charge our phones and continue planning the next day’s tasks.
One evening, while assisting a neighbour with an electrical fault, my close friend Matthew attempted to test for live and neutral wires. In the process, he was suddenly struck by a live wire and became unconscious immediately.
We panicked and searched for help, but the environment was extremely rural. There were no clear emergency services nearby, and people in the area were unsure of where the closest functional hospital was located.
After about 20 minutes of uncertainty, someone finally pointed us to a hospital. We drove for approximately 45 minutes in urgency, only to discover that the facility was not equipped for emergency cases—it was primarily a dental clinic. They redirected us again.
By the time we reached the correct hospital, which was ironically much closer than we originally thought (near a school area in Port Harcourt), significant time had already been lost.
The doctors did everything they could, but shortly after intervention began, we were informed that Matthew had passed away.
What stayed with me was not only the emotional loss, but the realization that delayed access caused by misinformation and lack of clear healthcare discovery systems contributed heavily to the outcome.
From that moment, I became deeply aware that in emergencies, information is as critical as treatment itself.
Who Benefits from AtlasRoute
AtlasRoute is designed for anyone who needs fast, reliable access to nearby healthcare facilities, especially in situations where time, location, or lack of local knowledge becomes a barrier to getting care. Its impact spans multiple real-world scenarios affecting millions of people globally:
1. Emergency Accident Cases
Individuals involved in road accidents, workplace injuries, or sudden medical emergencies who need to quickly locate the nearest accessible hospital or clinic. In critical moments, AtlasRoute reduces the delay between incident and care by providing immediate, relevant facility options.
2. Elderly or Vulnerable Individuals
Elderly users or vulnerable individuals who may not easily remember nearby healthcare locations and need simple, guided access to medical facilities during sudden illness or discomfort, without relying on others for directions.
3. Mobility Workers (Drivers, Delivery, Logistics)
People who spend long hours on the road such as truck drivers, delivery personnel, and rideshare operators who may experience fatigue, minor injuries, or sudden illness while traveling through unfamiliar areas and need quick access to nearby care.
4. Tourists and Foreign Visitors
International visitors or travelers unfamiliar with local infrastructure who need fast discovery of nearby hospitals or clinics. AtlasRoute helps reduce confusion caused by language barriers and lack of local knowledge.
5. Students in New Environments
Students living in new cities or campuses who may suddenly need medical assistance but are unfamiliar with nearby healthcare facilities and their locations.
6. Field and Remote Workers
Workers in construction sites, industrial zones, oil fields, and remote environments where immediate access to medical facilities is limited and knowing the nearest available care center is critical for safety.
7. Disaster or Disruption Scenarios
Users affected by floods, road closures, power outages, or other disruptions where normal access routes to hospitals may be blocked, requiring quick identification of alternative functioning healthcare facilities.
8. Underserved or Rural Populations
Communities in rural or underserved regions where healthcare facilities exist but are not widely known or easily discoverable, making it difficult for residents to access timely medical care.
AtlasRoute ensures that in all these scenarios, the gap between needing care and finding care is significantly reduced.
What it does
AtlasRoute is an AI-powered healthcare discovery system built to help users instantly locate the nearest and most relevant medical facilities in any situation.
It provides intelligent access to:
- Hospitals
- Clinics
- Pharmacies
- Dental facilities
- Emergency care points
The system is designed to work in a simple, step-by-step interaction flow that makes emergency access faster and more reliable:
The user enters a natural language request, such as “I need the nearest hospital” or “Find a clinic near me”.
The AI responds by generating a secure location permission link. The user is prompted to click the link and grant location access.
Once the user grants permission, they return to the chat and confirm that access has been approved.
The AI then uses the user’s real-time location to retrieve and analyze nearby healthcare facilities.
Finally, AtlasRoute returns a curated list of the most relevant nearby options, including travel distance, estimated time, and facility type—prioritizing what is most suitable for the user’s situation.
This approach ensures that users stay in control of their location data while enabling the system to deliver accurate, real-time healthcare discovery tailored to their exact position and needs.
This allows AtlasRoute to go beyond “nearest result” and instead provide the most appropriate nearby care option.
How we built it
AtlasRoute is built using a combination of geospatial infrastructure, healthcare data standards, and AI orchestration through the Model Context Protocol (MCP).
Core Technologies:
- OpenStreetMap (OSM): Global healthcare facility dataset
- Nominatim API: Geocoding and location resolution
- OpenRouteService: Route calculation and travel estimation
- Model Context Protocol (MCP): AI-to-tool communication layer
- FHI: Structured healthcare context for patient-aware recommendations
Architecture:
The system follows a tool-augmented AI model:
- User submits a natural language request (e.g., “I need the nearest hospital”)
- The AI interprets intent and urgency
- MCP selects the appropriate tool (hospital/clinic/pharmacy search)
- External APIs return real-world structured data
- Results are filtered, ranked, and returned with routing information
A key design decision was ensuring the AI never “hallucinates” locations. Every response is strictly grounded in verified geospatial APIs.
Challenges we ran into
1. Preventing hallucination in critical outputs
Since healthcare is safety-sensitive, one of the biggest challenges was ensuring the AI does not generate or guess facility data. We solved this by enforcing strict tool-based responses via MCP.
2. Inconsistent global healthcare data
Healthcare facility data varies significantly across regions. Some areas have detailed metadata, while rural regions often have incomplete or outdated information. We had to design fallback strategies for incomplete datasets.
3. Balancing simplicity and complexity
Users in emergency situations cannot deal with complex filters or multi-step searches. However, the backend system requires handling multiple data layers. We had to design a system that feels simple externally but remains powerful internally.
4. Latency constraints
In emergency scenarios, even small delays matter. We optimized API calls, reduced redundant processing, and streamlined AI-tool interaction flow to minimize response time.
Accomplishments that we're proud of
- Successfully integrating AI reasoning with real-world geospatial APIs using MCP
- Building a system that dynamically handles multiple healthcare facility types
- Designing an emergency-first interface that prioritizes speed and clarity
- Ensuring strict grounding of AI outputs in verified external data sources
- Transforming a personal real-life experience into a functional, scalable system
- Creating a flexible architecture that can extend globally
What we learned
This project reinforced a key principle: AI is most powerful when it is connected to real-world systems rather than operating in isolation.
We also learned:
- The importance of grounding AI outputs in verified data sources
- That healthcare accessibility is not just about infrastructure, but discoverability
- That system reliability is more important than feature complexity in life-critical applications
On a deeper level, we learned how technology can transform personal loss into meaningful impact when applied thoughtfully.
What's next for AtlasRoute
AtlasRoute is designed as a foundation for a broader emergency healthcare intelligence system.
Future improvements include:
1. Real-time hospital capacity awareness
Integrating live data on hospital availability, ICU capacity, and emergency load.
2. Enhanced facility classification
Expanding metadata to include trauma level, pediatric availability, specialist services, and emergency readiness.
3. Offline-first emergency mode
Allowing users to access cached healthcare location data in low-connectivity environments.
4. Global scaling and rural expansion
Improving data coverage in underserved and rural regions through better mapping and validation systems.
5. AI-driven triage assistance
Enhancing urgency detection to better distinguish between emergency, urgent, and non-urgent situations.
6. Emergency system integrations
Exploring integration with emergency response networks and public safety systems.
AtlasRoute is ultimately working toward a simple goal:
"No one should lose critical time trying to find care that should already be within reach"
Built With
- model-context-protocol-(mcp)
- nominatim-api
- openrouteservice
- openstreetmap-(osm)
- typescript
Log in or sign up for Devpost to join the conversation.