Being able to visualize protein localizations within cells and tissues by means of immuno-fluorescence microscopy has been key to developments in cell biology and beyond. Gut et al. present a high-throughput method that achieves the detection of more than 40 different proteins in biological samples across multiple spatial scales. This allows the simultaneous quantification of their expression levels in thousands of single cells; captures their detailed subcellular distribution to various compartments, organelles, and cellular structures within each of these single cells; and places all this information within a multicellular context. Such a scale-crossing dataset empowers artificial intelligence–based computer vision algorithms to achieve a comprehensive profiling of intracellular protein maps to measure their responses to different multicellular, cellular, and pharmacological contexts, and to reveal new cellular states.