This study investigates how and to what extent Animoji are vocally performed in private messages to recipients of varying levels of intimacy. Animoji users matched seven Animoji with seven relationships at different degrees of social distance and recorded a message for each pairing. The messages of 17 participants were narrowly transcribed, coded for performance features, and analyzed by Animoji and relationship type. Participants tended to modify their speech in different ways (e.g., via prosody, accent, lexis) with different Animoji, although the modifications were not as extreme as in previously studied public Animoji performances. Relationship type appears to be a better predictor of performativity than Animoji overall, but there are notable exceptions.