AI Agent-Enabled Development of Enhanced DataYee for Robust and Efficient AFM Force Spectroscopy Analysis

High-throughput atomic force microscopy-based single-molecule force spectroscopy (AFM-SMFS) produces large, heterogeneous data sets that require force−curve screening, polymer model fitting, molecular event annotation, and expert verification. These interconnected operations limit analytical throughput and can introduce uncertainty when noisy or nonideal trajectories are processed. Here, we report the expert-supervised, AI agent-enabled development of DataYee, an integrated platform for AFM-SMFS analysis. The previously developed platform is referred to as original DataYee, whereas the version developed in this study is referred to as enhanced DataYee. During development, the AI agent assisted with literature retrieval, codebase navigation, coordinated modification of multiple software modules, refactoring, and debugging. Researchers defined the analytical requirements and verified the physical models, parameter constraints, software behavior, and experimental outputs. Enhanced DataYee combines convolutional neural network-based prescreening with support vector machine classification using interpretable mechanical features, noise-adaptive fitting window selection, worm-like regime crossover-informed multi-start worm-like chain fitting, and domain annotation constrained jointly by contour-length increments and unfolding forces. Functional refinements further improve data import, curve management, exception handling, and routine operation. Evaluation using a representative AFM-SMFS data set showed that enhanced DataYee improved the fitting and retention of noisy or nonideal trajectories while substantially reducing the number of curves requiring expert review. Following targeted expert verification, the resulting unfolding-force distribution remained consistent with manual analysis while including more validated molecular events. Enhanced DataYee therefore provides an efficient and traceable expert-in-the-loop workflow for high-throughput SMFS analysis and demonstrates how AI agents can support the scientifically supervised development of laboratory analytical software.