Perancangan Sistem Backend Real-Time Berbasis FastAPI dan Firebase Untuk Klasifikasi Sampah Bahan Berbahaya dan Beracun Serta Pengambilan Lokasi Dropbox Pada Aplikasi Mobile TOXMAP

Gading Aurelia Nabila Lubis
Rita Purnamasari
Khaerudin Saleh

eProceedings of Engineering
Vol. 12 No. 4 (2025): Agustus 2025

Hazardous household waste (B3) poses serious risks to the environment and public health when improperly managed. To address this, TOXMAP was developed as a mobile-based system that combines real-time image classification with location-based waste disposal guidance. This study presents the backend development of TOXMAP, which integrates a FastAPI server to process image inputs and classify waste types using a pre-trained Support Vector Machine (SVM) model. Firebase services support user authentication, image storage, and dropbox location retrieval. To complement this backend, the TOXMAP frontend was built using Flutter, enabling cross-platform mobile deployment with support for camera access and real-time image submission. Integration and load testing show that the system can respond in under one second per request while maintaining classification accuracy and user responsiveness. This architecture offers a lightweight, scalable, and user-accessible infrastructure to promote public awareness and responsible household B3 waste management.Keywords— backend, real-time system, image classification, FastAPI, Firebase, B3 waste