Anticancer peptides (ACPs) are cationic amphiphiles that preferentially kill cancer cells through folding-dependent membrane disruption. Although ACPs represent attractive therapeutic candidates, particularly against drug-resistant cancers, their successful translation into clinical practice has gone unrealized due to their poor bioavailability, serum instability and, most importantly, severe hemolytic toxicity. Here, we exploit the membrane-specific interactions of ACPs to prepare a new class of peptide-lipid particle, we term a lipopeptisome (LP). This design sequesters loaded ACPs within a lipid lamellar corona to avoid contact with red blood cells and healthy tissues, while affording potent lytic destruction of cancer cells following LP-membrane fusion. Biophysical studies show ACPs rapidly fold at, and integrate into, liposomal membranes to form stable LPs with high loading efficiencies (>80%). Rational design of the particles to possess lipid combinations mimicking that of the aberrant cancer cell outer leaflet allows LPs to rapidly fuse with tumor cell membranes and afford localized assembly of loaded ACPs within the bilayer. This leads to preferential fusolytic killing of cancer cells with minimal collateral toxicity towards non-cancerous cells and erythrocytes, thereby imparting clinically relevant therapeutic indices to otherwise toxic ACPs. Thus, integration of ACPs into self-assembled LPs represents a new delivery strategy to improve the therapeutic utility of oncolytic agents, and suggests this technology may be added to targeted combinatorial approaches in precision medicine. STATEMENT OF SIGNIFICANCE: Despite their significant clinical potential, the therapeutic utility of many ACPs has been limited by their collateral hemolysis during administration. Leveraging the membrane-specific interactions of ACPs, here we prepare self-assembled peptide-lipid nanoparticles, or 'lipopeptisomes' (LPs), capable of preferentially fusing with and lysing cancer cell membranes. Key to this fusolytic action is the construction of LPs from lipids simulating the cancer cell outer leaflet. This design recruits the oncolytic peptide payload into the carrier lamella and allows for selective destruction of cancer cells without disrupting healthy cells. Consequently, LPs impart clinically relevant therapeutic indexes to previously toxic ACPs, and thus open new opportunities to improve the clinical translation of oncolytics challenged by narrow therapeutic windows.