Masses, radii and luminosities of distant stars can only be measured accurately in eclipsing binaries. The most massive eclipsing binary currently known is a Wolf-Rayet binary (WR20a), which consists of two ~80 Mo stars in a 3.7 day orbit. Analogs of WR20a are bound to exist both in massive stellar clusters in our Galaxy and in nearby galaxies. The nearest ones are located in the young massive clusters near the Galactic Center: the Center, Arches, and Quintuplet clusters and in Galactic super star clusters, such as Westerlund 1 and NGC 3603. The severe amount of reddening in the Galactic disk makes the study of Galactic clusters challenging. However, with current 8-m class telescopes, the study of massive stars in nearby galaxies is also feasible. The nearest Local Group galaxies (LMC, SMC, M31, M33) provide the perfect laboratory for studying massive stars and determining their properties as a function of metallicity. I will present the first results of a survey of the most massive stars in eclipsing binaries in the Milky Way and in the Local Group. Measurements of their fundamental parameters will constrain untested formation and evolution models, confirm the dependence of evolution on metallicity, probe the upper stellar mass limit and provide insight on the rate and nature of supernovae and gamma-ray bursts.