Nanolimes have been commercially available for over a decade as a remineralization agent for natural stone to combat deterioration. While they have been applied successfully and studied extensively, their penetration abilities in different materials have not yet been readily quantifiable in situ and in real time. Using two transparent pore-imitating test systems (acrylic glass (PMMA) and polydimethylsiloxane (PDMS)) and light microscopy, the penetration coefficients (PCs) of two nanolimes (CaLoSiL (CLS) and Nanorestore Plus (NRP)), as well as their solvents, were determined experimentally in square channels of about 100 µm diameter. Their PCs and those for a previously published glass–resin-based test system were also predicted based on measurable material parameters or literature values using the Lucas–Washburn equation. Additionally, a liquid mineral precursor (LMP) of calcium carbonate based on complex coacervation (CC) was investigated as an alternative to the solid particle dispersions of nanolime. In general, the dispersions behaved like their pure solvents. Overall, trends could be reasonably well predicted with both literature and experimentally determined properties using the Lucas–Washburn equation. In absolute terms, the prediction of observed infiltration behavior was satisfactory for alcohols and nanolimes but deviated substantially for water and the aqueous LMP. The commercially available PMMA chips and newly designed PDMS devices were mostly superior to the previously published glass–resin-based test system, except for the long-term monitoring of material deposition. Lastly, the transfer of results from these investigated systems to a different, nontransparent mineral, calcite, yielded similar PC values independently of the original data when used as the basis for the conversion (all PC types and all material/liquid combinations except aqueous solutions in PDMS devices). This knowledge can be used to improve the targeted design of tailor-made remineralization treatments for different application cases by guiding solvent choice, and to reduce destructive sampling by providing a micromodel for pretesting, if transferability to real stone samples proves demonstrable in the future.