Learned visual attributes have become an important feature for recognition and search. Visual Attributes can be computed using SVMs, which then require normalization and fusion for effective searching. In this demo, we present a MugHunt, a highly scalable Google-like face search system based around Meta-Recognition fusion and implemented using of Cassandra, allowing searching over millions of face images using attribute descriptions (gender, race, hair color, expression, etc.). This demo will also include a look under the covers at the Meta-Recognition based normalization and implementation of similarity searching.