Result: Machine learning-assisted sensing platform for simultaneous and visual detection of tetracycline hydrochloride and ofloxacin.

Antibiotics are essential in modern medicine; however, their overuse and improper disposal have caused significant environmental contamination, which impacts aquatic systems, food chains, and ecosystems while posing serious risks to human health. Current antibiotic detection methods are typically confined to laboratories and require professional expertise, creating a gap for portable sensors suitable for everyday use. In this study, we developed a machine learning (ML)-assisted visual sensing platform for the dual detection of tetracycline hydrochloride (Tet) and ofloxacin (Ofl). Fluorescent dyes Rhodamine B (RhB) and Fluorescein (Flu) were efficiently immobilized on water-stable nano-UiO-67 via post-synthetic modification. This material was further integrated with sodium alginate (SA) to construct a fluorescence-based sensor, enabling visual quantification of Tet and Ofl across dynamic concentration ranges (0–600 μM). The platform demonstrated robust performance in complex matrices—including lake water, tap water, and milk—achieving recoveries of 99.38–105.72 %, thereby validating its utility for real-world antibiotic monitoring. Critically, rapid identification and concentration determination of Tet and Ofl were achieved through adaptive learning of image features using a Python-based smart system, the RGB-ExponentialFit Analyzer (RGB-EFA). Combining metal-organic framework (MOF)-based fluorescence sensing with the RGB-EFA ML algorithm establishes a novel paradigm for antibiotic detection: rapid, cost-effective, operationally simple, portable, and visually interpretable. This work underscores the transformative potential of ML-enhanced biosensors in addressing antibiotic pollution and advancing point-of-need diagnostics. [ABSTRACT FROM AUTHOR]