A typical arc-synthesis generates many types of fullerenes and endohedrals. Resulting soot extracts contain a complex mixture of >50 types of fullerenes, metallofullerenes, and their structural isomers. Prior to application development, novel separation methods are required to fractionate this rich array of metallic, metallic carbide, metallic nitride, and metallic oxide endohedrals, all of which can be present in a single, soot extract. Herein, we report the discovery of CuCl2 as a Lewis acid that will selectively precipitate only the more reactive members of each of these endohedral families. The more reactive Sc4O2@I-h-C-80, Sc3C2@I-h-C-80, and Sc3N@D-3h-C-78 endohedrals are quickly removed from extracts to greatly decrease the number of endohedrals present in a sample. Experiments indicate that enrichment factors of several orders of magnitude can be achieved within minutes of reaction time. CuCl2 also has sufficient selectivity to resolve and separate structural isomers, as demonstrated with Er-2@C-82 (isomer I, C-s(6)-C-82 versus isomer III). The selective complexation of CuCl2 with fullerenes can be correlated to their first oxidation potential. We estimate a significantly lower threshold of precipitation for CuCl2 (<0.19 V) compared to stronger Lewis acids. Fullerenes and metallofullerenes having first oxidation potentials above 0.19 V tend to remain unreacted in solution. In contrast, species with first oxidation potentials below 0.19 V (vs Fc/Fc(+)) precipitate via complexation, and are easily decomplexed. CuCl2 is compared to Lewis acids having higher precipitation thresholds (e.g., FeCl3) in our goal to predict a priori which endohedrals would remain in solution versus which endohedral species would complex and precipitate. The ability to predict endohedral precipitation a priori is beneficial to the design of purification strategies for metallofullerenes.