BACKGROUND: Disease registries assist with clinical practice improvement. The Australian Stroke Clinical Registry aims to provide national, prospective, systematic data on processes and outcomes for stroke. We describe the methods of establishment and initial experience of operation. METHODS: Australian Stroke Clinical Registry conforms to new national operating principles and technical standards for clinical quality registers. Features include: online data capture from acute public and private hospital sites; opt-out consent; expert consensus agreed core minimum dataset with standard definitions; outcomes assessed at 3 months poststroke; formal governance oversight; and formative evaluations for improvements. RESULTS: Qualitative feedback from sites indicates that the web-tool is simple to use and the user manuals, data dictionary, and training are appropriate. However, sites desire automated data-entry methods for routine demography variables and the opt-out consent protocol has sometimes been problematic. Data from 204 patients (median age 71 years, 54% males, 60% Australian) were collected from four pilot hospitals from June to October 2009 (mean, 50 cases per month) including ischaemic stroke (in 72%), intracerebral haemorrhage (16%), transient ischaemic attack (9%), and undetermined (3%), with only one case opting out. CONCLUSION: Australian Stroke Clinical Registry has been well established, but further refinements and broad roll-out are required before realising its potential of improving patient care through clinician feedback and allowance of local, national, and international comparative data.