A large number of therapeutically useful cyclic and linear peptides of bacteria or fungal origin are synthesized via a template-directed, nucleic-acid-independent nonribosomal mechanism. This process is carried out by mega-enzymes called nonribosomal peptide synthetases (NRPSs). NRPSs contain repeated coordinated groups of active sites called modules, and each module is composed of several domains with different catalytic activities. The familiarity to these domains lays base for the future genetic engineering of NRPSs to generate entirely "unnature" products. The details about NRPSs domain structures and the exploitation of NRPSs are described in this review.