An end-to-end, AI-powered travel planning mobile application delivering dynamic, personalized Saudi itineraries.
Travelers exploring Saudi regions often face fragmented information sources and static, generic recommendations that fail to factor in real-time user interest vectors, trip duration limits, and geographic constraints.
Terhal bridges this gap by combining an intelligent recommendation engine with a seamless cross-platform mobile experience.
- Frontend: Cross-platform mobile client built with Flutter for intuitive user onboarding and itinerary visualization.
- Backend API: Asynchronous, high-throughput REST APIs built using FastAPI and deployed on Render.
- AI Core: Constraint-aware recommendation engine mapping dynamic user preference vectors to localized destination features.
- Database & Auth: Scalable cloud document storage using MongoDB combined with Firebase for authentication and session state management.
| Domain | Technologies |
|---|---|
| Mobile Client | Flutter, Dart |
| Backend Framework | FastAPI (Python), Uvicorn |
| AI / Machine Learning | Python, Scikit-learn, Pandas, Vector Similarity |
| Databases | MongoDB Atlas, Firebase |
| Deployment & DevOps | Render, Git / GitHub |
- Vector-Based Personalization: Dynamically scores destinations based on personalized user preferences and travel styles.
- Constraint-Aware Scheduling: Generates optimized itineraries taking into account user time limits and destination operating constraints.
- Rich Regional Catalog: Curated, structured dataset of tourist attractions, cultural sites, and dining experiences across Saudi Arabia.
- Real-Time Synchronization: Fast API communication ensuring responsive updates across client sessions.
├── ai/ # Recommendation algorithms & vector similarity models
├── mobile/ # Flutter mobile application codebase
├── models/ # Data schemas and Pydantic validation models
├── routes/ # FastAPI endpoint definitions & controllers
└── web/ # Web dashboards and auxiliary interfaces