Redox homeostasis is a key factor in maintaining cellular function and health. The main determinant of the intracellular redox potential is ubiquitous glutathione (GSH) together with its oxidized dimer (GSSG). Importantly, both redox equilibrium and GSH homeostasis are highly compartmentalized. Thus, tools with subcellular resolution are required to gain a detailed understanding of cellular redox chemistry, as well as the related signaling pathways and stress responses. In this thesis, we present a novel GSH sensor for live-cell imaging in mammalian cells. By combining a spirocyclizing, GSH-sensitive synthetic fluorophore and the self-labeling protein HaloTag (HT), we created a sensor with a fluorescence and reactivity turn-on mechanism. Structural biology revealed the ligand-protein interactions that facilitate this unique feature. Localized expression of HT as a fluorescent fusion protein enabled straightforward targeting and a robust ratiometric read-out. In live-cell experiments, this sensor was able to report reliably on changes in subcellular GSH. Calibration of the sensor allowed the determination of the absolute GSH concentration in various organelles. Our results clearly suggested separated GSH pools in the cytosol, the endoplasmic reticulum (ER) and the nucleus. A second-generation sensor ought to improve slow equilibration kinetics with GSH and finetune the dynamic range for organelles with low GSH concentrations. Therefore, we established a pipeline for the generation of site-saturation mutagenesis libraries to screen for HT mutants that enhance the desired characteristics. Besides measuring GSH, another goal of this work was the controlled and targeted manipulation of the GSH/GSSG equilibrium. As redox homeostasis and stress responses are compartment-specific, to probe the cellular reaction to redox perturbations, tools with subcellular selectivity are needed too. In this context, reductive stress is way less explored than oxidative stress but could be especially interesting in organelles with an oxidizing environment like the ER. To probe subcellular reductive stress, we wanted to create a stress inducer that could release a phosphine upon binding to HT. The released phosphine could reduce GSSG, hence impacting redox homeostasis. Again, by expressing HT in the compartment of interest, targeted activation can be achieved. Successful release is indicated by fluorescence turn-on if the phosphine is carried by a fluorogenic scaffold. We considered rhodamine-, hybrid coumarin- and cyanine-based probes and different substitution patterns were tested to obtain a stable phosphonium that is selectively released. But despite promising preliminary results, these probes still need further optimization to become a robust tool for probing reductive stress in living cells. Lastly, we explored how multiplexed fluorescence lifetime imaging with self-labeling proteins can be used to observe several subcellular compartments concurrently. We adapted the environment sensitive COUPY dye scaffold to bind HT. Even subtle changes such as mutations in the protein affected the local environment of the dye, which was reflected in altered fluorescence lifetimes. A screening of only 10 mutants identified two versions of HT, which yielded separable signals by lifetime unmixing. With the unbound dye accumulating in mitochondria and the HT variants being expressed in the nucleus and the Golgi apparatus, three organelles could be simultaneously observed.