The quantification of carbon in silicate materials, such as volcanic glasses and experimental run products, is crucial for understanding magmatic processes, volatile behavior, and degassing mechanisms. Electron probe microanalysis (EPMA) offers high spatial resolution and low detection limits, making it a valuable tool for analyzing carbon at micrometer scales. However, accurate carbon measurement via EPMA presents challenges due to issues such as contamination, beam-induced carbon deposition, and matrix effects. This study evaluates the reliability of carbon analysis in silicate matrices using wavelength-dispersive spectrometry (WDS) on an electron microprobe. We investigate the effects of sample preparation, coating methods, and analytical conditions on carbon detection. A series of well-characterized materials, including natural and synthetic glasses ranging from basaltic to felsic and alkaline compositions, were used to assess detection limits and precision. The results indicate that carbon contamination from the environment can be minimized by optimizing vacuum conditions and pre-analysis cleaning procedures. Additionally, we compare the effect of different analytical settings, such as beam current and accelerating voltage. At 10 kV and 60 nA, detection limits for a single analysis spot were found to be <800 ppm. Our findings highlight the potential of EPMA for in situ carbon quantification in silicate glasses when appropriate sample preparation and calibration methods are applied. The study provides insights into best practices for reducing artifacts and improving the accuracy of carbon measurements, contributing to a better understanding of magmatic carbon storage and release.