Two international scientific publishers collaborated to develop an Online Submission and Peer Review System (OSPREY) for their journals. Our goals were to meet market demand, increase editorial efficiency and streamline the transition from peer review to publishing. One of the publishers (National Research Council (NRC) Research Press, Canada) had previously purchased a third-party system that was subsequently discontinued by the vendor. Because of this experience and its complex requirements, we decided to build rather than buy a new system. The collaboration with the second publisher, Commonwealth Scientific and Industrial Research Organisation (CSIRO) Publishing, Australia, allowed sharing of resources within a common vision and goals. Agile development through the use of iterations allowed us to continuously add functionality, make improvements and incorporate new requirements. The development team included technical staff as well as stakeholders, future users, business analysts and project managers. The architecture chosen was based on open source technologies, with Java servlets and Java Server Pages for the Web interface. OSPREY currently supports 32 journals at the two publishers. Users accomplish all regular tasks in peer review (submission, selection and invitation of reviewers, submission of review, recommendations and decision) through the software. Editorial staff verifies submissions, sends correspondence and assigns customizable roles and tasks. All tasks are accomplished through a Web browser accessing the application on central servers at the publisher, with no special software or configuration required for any users. Currently, the system integrates with the publishing system by generating manuscript metadata in an XML format, although closer integration with a workflow management system is planned. Since OSPREY implementation, the number of submissions has risen, although marketing and higher ranking of the journals are also factors. For the future, we plan to add new functionality for business tasks and for parsing, tagging and linking of article references.