Over the past decade, several classes and/or prototypes of cell penetrating peptides (CPP) have been identified and investigated in multiple aspects. CPP represent peptides, which show the ability to cross the plasma membrane of mammalian cells, and may thus give rise to the intracellular delivery of problematic therapeutic cargos, such as peptides, proteins, oligonucleotides, plasmids and even nanometer-sized particles, which otherwise cannot cross the plasma membrane. Most of the currently recognized CPP are of cationic nature and derived from viral, insect or mammalian proteins endowed with membrane translocation properties. The exact mechanisms underlying the translocation of CPP across the cellular membrane are still poorly understood. However, several similarities in translocation can be found. Early studies on CPP translocation mechanisms tended to suggest that the internalization of these peptides was neither significantly inhibited by low temperature, depletion of the cellular adenosine triphosphate (ATP) pool, nor by inhibitors of endocytosis. Moreover, chemical modification of the peptide sequence, such as the synthesis of retro-, enantio- or retroenantio-analogs, appeared not to affect the internalization properties. Therefore, translocation was concluded to result from direct, physical transfer through the lipid bilayer of the cell membrane. Later studies, however, showed convincing evidence for the involvement of endocytosis as the dominating mechanism for cellular internalization. In addition to describing the general properties of the commonly recognized classes of CPP, in this review we will also point out some limitations and typical pitfalls of CPP as carriers for therapeutics. In particular we will comment on emerging discrepancies with the current dogma, on cell-to-cell variability, biological barrier permeability, metabolic fate, toxicity and immunogenicity of CPP.