Air pollution prediction services are often locked behind proprietary dashboards, limiting their reuse in external applications and city-scale decision support tools. This paper presents a city-agnostic air pollution prediction system that combines an H3 hexagonal spatial index with a RESTful data API and an embeddable web widget for public access. The backend integrates measurements from heterogeneous fixed and mobile sensors, stores time series in a document database, and maintains current aggregations and predictions in a relational schema linked to hexagonal cells. Prediction results and raw measurements are exposed through JSON endpoints, returning both sensor-level data and hex-level aggregations (N \(_{2}\) , PM \(_{10}\) , PM \(_{2.5}\) , temperature, humidity, pressure, etc.). In addition to this API, a configurable JavaScript widget renders interactive hexagon maps and meter views that can be embedded into any HTML5 website to inform citizens, drivers, and third-party services in real-time. The system is designed to be reproducible and portable: any city with basic sensor infrastructure can deploy the stack, expose open data, and integrate its predictions into existing web portals, mobile apps, and transport information systems.