A system and method is provided that dynamically creates, manipulates and transfers digital images over a communication network on a per request basis in an efficient manner, wherein the composition, format, presentation, and content of the digital images are determined by directives presented at the time of the request. The system includes a parser that takes a request and interprets it into a set of individualized operation commands that will, in sequence, be performed to create or alter an image. The commands derived from the parser are defined as a "job. One or more image processing engines may be called upon as the job is processed to perform requested image transformations. An output formatter takes the resulting processed image and converts it to the desired format stated in the request.