This thesis focuses on indexing structures for similarity searching. The main objective is to design and implement an index based on metric hulls. A metric hull is a recently-proposed method for a compact representation of a set of metric data objects. The prototype implementation will provide kNN search and approximate kNN search operations. Experimental evaluation on a real-life dataset will be done and the results will be compared to ones of the M-tree. The text part of the thesis will describe the principle of metric hulls, the index structure created, and the commentary on the results of the experimental evaluation.