<p>Auxetic metamaterials, valued for their lateral expansion under tension, are attractive for applications from wearable protection to biomedical implants. Conventional workflows that rely solely on finite-element optimisation hinge on repeated calibration of material and contact models and often under-represent fabrication-induced deviations. To address these challenges, we present Clip4AMO (Closed-loop Intelligent Platform for Auxetic Metamaterial Optimisation), a low-barrier workflow assembled from retrofitted consumer-grade modules that supports unattended design–fabricate–test–update rounds driven by surrogate-assisted sequential experimental design. Candidate selection uses a random-forest surrogate and a lower confidence bound criterion to efficiently explore design spaces within limited experimental iterations. In a thermoplastic polyurethane case study on an eight-parameter chiral unit-cell family, Clip4AMO reaches a best Poisson’s ratio of approximately − 0.75 within 11 rounds while yielding multiple near-optimal alternatives. Leveraging Shapley Additive Explanations, we clarify how key geometric parameters influence auxetic behaviour, offering transferable insights for targeted design. Repeat tests and Gaussian-process uncertainty propagation under ± 5% parameter perturbations indicate that champion designs retain strongly negative Poisson’s ratios under plausible geometric variability. Overall, Clip4AMO offers a data-driven, scalable and interpretable route for optimising multi-parameter deformable structures, and it can be extended to multi-objective tasks such as energy absorption and fatigue-life enhancement.</p>