The incorporation of transition metal catalysts to the bioorthogonal toolbox has opened the possibility of producing supra-stoichiometric amounts of xenobiotics in living systems in a non-enzymatic fashion. For medical use, such metals could be embedded in implantable devices (i.e. heterogeneous catalyst) to "synthesize" drugs in desired locations (e.g. in a tumour) with high specificity and for extended periods of time, overcoming the useful life limitations of current local therapy modalities directed to specific organ sites (e.g. brachytherapy, controlled release systems). To translate this approach into a bona fide therapeutic option, it is essential to develop clinically-accessible implantation procedures and to understand and validate the activation process in relevant preclinical models. Herein we report the development of a novel Pd-activatable precursor of the red-fluorescent drug doxorubicin and Pd devices of optimized size and activity. Screening in state-of-the-art cancer models provided fundamental insights into the insertion protocols, safety and stability of the devices and into the prodrug distribution profile before and after activation.