Admetica is an open-source batteries-included toolkit for predicting pharmacokinetic (ADMET) properties of small molecules from SMILES strings. It bundles 22 pre-trained Chemprop (message-passing neural network) models covering Absorption, Distribution, Metabolism, Excretion, and Toxicity endpoints, along with a CLI (`admetica_predict`) and a Dockerised Flask/Swagger REST API (`POST /predict`) for batched CSV-based inference.