In this paper, self-recorded interactions from two studies of verbal interaction in Denmark are analysed using key concepts from Erving Goffman's frame analysis and Don Zimmerman's theory of identities in talk. It is documented how the recorder is integrated into everyday interaction as soon as the informant in his or her capacity as temporary fieldworker has explained the reason for its being there and has obtained the permission from his or her partners to record them. In several excerpts ascribed identities and using the recorder as an addressee are used to create amusement and to exploit the possibilities for supplying the innocent overhearing researcher with new and shocking information. The question of language norms during recorded sessions was raised in conversations in both studies and the limits of privacy were brought up.