Flash™ is experiencing a breathtaking growth and has become one of the prevailing media formats on the Web. Unfortunately, no research effort has been dedicated to automatic retrieval of Flash movies by content, which is critical to the utilization of the enormous Flash resource. A close examination reveals that the intrinsic complexity of a Flash movie, including its heterogeneous components, its dynamic nature, and user interactivity, makes Flash retrieval a host of research issues. As the first endeavor in this area, we propose a generic framework termed as FLAME (FLash Access and Management Environment) embodying a 3-tier architecture that addresses the representation, indexing, and retrieval of Flash movies by mining and understanding of movie content. In particular, FLAME features a unique multi-level indexing and retrieval approach that supports characterization and retrieval of Flash at (1) object level, which describes the heterogeneous components embedded in a movie, (2) event level, which depicts the movie's dynamic effects constituted by the spatio-temporal features of objects, and (3) interaction level, which models the relationships between user behaviors and the consequential events. An experimental prototype for Web-based Flash retrieval is implemented to verify the feasibility and effectiveness of FLAME.