AI-integrated smartphone platform enables POC dual-channel glucose monitoring with an indicator-free nanozyme gel kit.

We developed an indicator-free, dual-mode glucose monitoring platform by synthesizing glucose oxidase (GOX)-templated MnO2 nanosheets via a microwave-assisted static biomineralization strategy. This approach preserves the enzyme's native activity and intrinsic fluorescence while forming a stable nanostructure. Mechanistically, the MnO2 nanosheets initially quench the GOX fluorescence via Förster Resonance Energy Transfer (FRET). In the presence of glucose, the enzymatically generated H2O2 reduces MnO2 to Mn2+, triggering simultaneous color fading and fluorescence recovery. The solution-based sensor exhibited a linear response from 0.005 to 10 mM with a detection limit (LOD) of 2.9 μM. To facilitate point-of-care testing (POCT), we embedded the nanomaterials into an agarose gel to create a "microtube lab". Integrated with a smartphone application based on deep learning algorithms (SmartGDS), the platform enables rapid, fluorescence-based and colorimetric quantification of salivary glucose within 30 min. The integrated system achieved a linear range of 0.01-10 mM with an LOD of 8.8 μM. Validation with clinical saliva samples demonstrated high accuracy comparable to commercial blood glucose kits, validating the platform's potential for non-invasive, low-cost health management.