WhatsApp has remained under the radar for it is scarcely accessible to overt scholarly scrutiny. Encrypted chat apps allow for a certain degree of perceived secrecy. Yet the high frequency of civic engagement makes ethnographic research a time-consuming exercise. This article investigates how digital ethnography inside WhatsApp groups requires up-to-date, innovative ethical guidelines. We suggest a two-pronged approach. On the one hand, we should rethink and update 'known' ways of doing ethics, undertaking at least three conceptual operations: going back to the basics, positing as central the notion of 'do not harm', which allows to recentre the user within the research process; avoid reducing research ethics to a one-stop checklist, to privilege instead a recursive, iterative and dialogic process able to engage research subjects; moving past the consent form as the sole and merely regulatory moment of the researcher-research subject relationship. On the other hand, while thinking through innovative ways of considering ethics in chat app research, we ought to take infrastructure seriously, both the site of research and the research ecosystem; embrace transparency and avoid by all means covert bypasses; and guarantee full anonymisation to our research subjects.