Aims: Community-based residential alternatives to hospitalization are an emerging service model. Evidence for their acceptability and effectiveness is promising but limited. Prevention and Recovery Care (PARC) services are one such residential model, offering short-term subacute treatment and care (usually between 7 and 28 days). PARC services in Victoria, Australia, are designed to support consumers with severe mental illness to either avoid a psychiatric hospital admission (step-up care) or transition from hospital back into the community (step-down care). As a precursor to a series of studies investigating the appropriateness, effectiveness and efficiency of PARC services, we aimed to investigate whether a typology of PARC services can be developed. Methods: A manager or other appropriately knowledgeable staff member from each of the 19 adult PARC services included in the study completed a tool based on PARC operational guidelines (the Victorian PARC service mapping questionnaire) and a validated instrument measuring the quality of care in residential mental health settings (the Quality Indicator for Rehabilitative Care, QuIRC). Thirty (of 42) stakeholders participated in a modified Delphi study to select 23 from the available 230 variables for entry into a hierarchical cluster analysis. Results: Cluster analysis produced three clusters of equal dissimilarity. At the 90% confidence level, there were four variables which were significantly different between clusters. These were the year the PARC was opened, the QuIRC Living Environment domain score, the proportion of all admissions that were a step-down admission from an inpatient unit, and how often families were invited to care meetings. Sensitivity analyses suggested the findings were robust to the method used to identify clusters. Conclusions: Although PARC services were broadly similar, their identified differences suggest there is variable model implementation across Victoria sufficient to generate a PARC service typology. This typology may prove important for interpreting differences in outcomes experienced by consumers and carers using PARC services, when applied in our analyses of service effectiveness. The value of conducting service mapping and typology studies is underscored. Further research to characterize subacute residential services, including recovery-promoting features of the built environment, is warranted.