A computer-implemented method of object linking from captured media records media (audio or video) from a scene of a program (e.g., TV show, movie or other similar audio visual content), identifies the program and the program scene from a watermark embedded in the captured media or by fingerprinting the captured media and matching the fingerprint to determined fingerprints in a database that relates determined fingerprints to programs and program scenes, determines participants in scene from a participant database that relates participants to programs and program scenes, generates links to objects and online commerce sites offering the linked objects. Records for the objects are contained in a virtual repository, a database that relates users (e.g., participants) to object records.