A user device's browser receives a page from a publishing system and presents content of the page within a window. Instructions from a media hosting system cause a frame to be created within the window. The frame includes a media upload object from a media hosting system that has a different address domain that the publishing system. A communication channel is established between the window and the frame to allow messages to be exchanged. The media upload object allows a user to create a media item and upload it to the media hosting system. Throughout the process of creating and uploading the media item, the media upload object transmits one or more messages to the page via the communication channel. Based on the messages, the page responds to what is occurring within the frame.