Oncolytic viruses infect, replicate in and kill cancer cells. HSV has emerged as a most promising candidate because it exerts a generally moderate pathogenicity in humans; it is amenable to attenuation and tropism retargeting; the ample genome provides space for heterologous genes; specific antiviral therapy is available in a worst case scenario. The first strategy to convert HSV into an oncolytic agent consisted in deletion of the γ134.5 gene which counteracts the protein kinase R (PKR) response, and of the UL39 gene which encodes the large ribonucleotide reductase subunit. Tumor specificity resided in low PKR activity, and high deoxyribonucleotides content of cancer cells. These highly attenuated viruses have been and presently are in clinical trials with encouraging results. The preferred route of administration has been intratumor or in tissues adjacent to resected tumors. Although the general population has a high seroprevalence of antibodies to HSV, studies in animals and humans demonstrate that prior immunity is not an obstacle to systemic routes of administration, and that oncolytic HSV (o-HSVs) do populate tumors. As the attenuated viruses undergo clinical experimentation, the research pipeline is developing novel, more potent and highly tumor-specific o-HSVs. These include viruses which overcome tumor heterogeneity in PKR level by insertion of anti-PKR genes, viruses which reinforce the host tumor clearance capacity by encoding immune cytokines (IL-12 or granulocyte-macrophage colony-stimulating factor), and non-attenuated viruses fully retargeted to tumor specific receptors. A strategy to generate o-HSVs fully retargeted to human epidermal growth factor receptor-2 (HER-2) or other cancer-specific surface receptors is detailed. Copyright © 2011 John Wiley & Sons, Ltd.