The screening and diagnosis of breast cancer is a major public health issue. Although deep learning models are proving highly effective in breast imaging, these models are not yet readily accessible to a wide audience. In order to promote the widespread dissemination of such models, this article introduces a free and open-source, integrated platform for the automated detection of masses on mammograms. A state-of-the-art RetinaNet model is trained on this task and the results of the inference are encoded using the DICOM-SR interoperable format. These contributions present a significant step towards overcoming the accessibility gap in deep learning for breast imaging.