CAT-S is a bioorthogonal photocatalytic proximity labeling method that enables in situ profiling of mitochondrial proteomes in a wide range of primary living samples. This technique uses a thioquinone methide (thioQM) labeling warhead and targeted bioorthogonal photocatalytic chemistry to label mitochondrial proteins in living cells with high efficiency and specificity. CAT-S was applied to diverse cell cultures, dissociated mouse tissues, and primary T cells from human blood, revealing native-state mitochondrial proteomic characteristics and uncovering previously unknown mitochondrial proteins such as PTPN1, SLC35A4 uORF, and TRABD. It also allows quantification of proteomic perturbations in dysfunctional tissues, such as diabetic mouse kidneys, revealing alterations in lipid metabolism that may drive disease progression. CAT-S offers non-genetic operation, generality, and spatiotemporal resolution, making it a promising tool for subcellular proteomic investigations of primary samples that are otherwise inaccessible. The CAT-S system was developed by tuning the reactivity of QM probes and introducing thioQM as a more efficient and accurate labeling warhead. This system enables the in situ labeling of mitochondrial proteins in living cells without the need for genetic manipulation. The system was validated using various cell lines, including HeLa, HEK293T, and K562 cells, and was shown to capture mitochondrial proteomes with high specificity and efficiency. The system was also applied to primary cells from living tissues, such as mouse kidney and spleen, revealing mitochondrial proteomes with high coverage and specificity. Additionally, CAT-S was used to profile mitochondrial proteomes in diabetic mouse kidneys, revealing major alterations in lipid metabolism pathways, with metabolic enzymes such as Cpt1b, Acsm2, and Aldh3a2 differentially regulated, suggesting that the obese-diabetic condition suppresses the detoxifying lipid metabolic machinery, driving the progression of kidney disorders. CAT-S was further applied to primary T cells from human peripheral blood, revealing mitochondrial proteomes with high specificity and efficiency. The system was shown to capture mitochondrial proteomes with high coverage and specificity, and to reveal major pathways involved in mitochondrial function, such as cellular respiration and mitochondrion organization. The system was also shown to be biocompatible, with no obvious cellular toxicity under experimental conditions, and to be flexible in terms of external light control, enabling precise control of labeling time-window and high temporal resolution for potential dynamic investigations. The CAT-S system offers several advantages as a state-of-the-art in situ mitochondrial proteomics technique, including the ability to capture mitochondrial proteomes in diverse living cell samples, including hard-to-transfect cells, primary cells from tissues and blood samples that are otherwise inaccessible by genetic engineering and enzymatic approaches. It also significantly enhances the capture efficiency and specificity over previous chemical methods, and shows high biocompatibility with no obvious cellular toxicity under experimental conditions. The system is flexibly controlled by externalCAT-S is a bioorthogonal photocatalytic proximity labeling method that enables in situ profiling of mitochondrial proteomes in a wide range of primary living samples. This technique uses a thioquinone methide (thioQM) labeling warhead and targeted bioorthogonal photocatalytic chemistry to label mitochondrial proteins in living cells with high efficiency and specificity. CAT-S was applied to diverse cell cultures, dissociated mouse tissues, and primary T cells from human blood, revealing native-state mitochondrial proteomic characteristics and uncovering previously unknown mitochondrial proteins such as PTPN1, SLC35A4 uORF, and TRABD. It also allows quantification of proteomic perturbations in dysfunctional tissues, such as diabetic mouse kidneys, revealing alterations in lipid metabolism that may drive disease progression. CAT-S offers non-genetic operation, generality, and spatiotemporal resolution, making it a promising tool for subcellular proteomic investigations of primary samples that are otherwise inaccessible. The CAT-S system was developed by tuning the reactivity of QM probes and introducing thioQM as a more efficient and accurate labeling warhead. This system enables the in situ labeling of mitochondrial proteins in living cells without the need for genetic manipulation. The system was validated using various cell lines, including HeLa, HEK293T, and K562 cells, and was shown to capture mitochondrial proteomes with high specificity and efficiency. The system was also applied to primary cells from living tissues, such as mouse kidney and spleen, revealing mitochondrial proteomes with high coverage and specificity. Additionally, CAT-S was used to profile mitochondrial proteomes in diabetic mouse kidneys, revealing major alterations in lipid metabolism pathways, with metabolic enzymes such as Cpt1b, Acsm2, and Aldh3a2 differentially regulated, suggesting that the obese-diabetic condition suppresses the detoxifying lipid metabolic machinery, driving the progression of kidney disorders. CAT-S was further applied to primary T cells from human peripheral blood, revealing mitochondrial proteomes with high specificity and efficiency. The system was shown to capture mitochondrial proteomes with high coverage and specificity, and to reveal major pathways involved in mitochondrial function, such as cellular respiration and mitochondrion organization. The system was also shown to be biocompatible, with no obvious cellular toxicity under experimental conditions, and to be flexible in terms of external light control, enabling precise control of labeling time-window and high temporal resolution for potential dynamic investigations. The CAT-S system offers several advantages as a state-of-the-art in situ mitochondrial proteomics technique, including the ability to capture mitochondrial proteomes in diverse living cell samples, including hard-to-transfect cells, primary cells from tissues and blood samples that are otherwise inaccessible by genetic engineering and enzymatic approaches. It also significantly enhances the capture efficiency and specificity over previous chemical methods, and shows high biocompatibility with no obvious cellular toxicity under experimental conditions. The system is flexibly controlled by external