Safety planning is an intervention that demonstrated to be successful in help individuals self-manage suicidal crises. The SERO suicide prevention app supports this safety planning in a digital manner. The objective of this paper is to identify the specific components of safety plans and thereby guiding the design of digital solutions in a way that it supports individuals’ needs. We analysed the safety plans of all 1848 at risk users registered with the SERO app by applying BERTopic and Google’s Gemma-2 Large Language Model to identify themes and topics mentioned in the safety plans. The analysis of the safety plans revealed that users frequently cited family and loved ones as motivations to live, identified certain physical signs and emotions as warning signs, and reported social contact and activities as coping strategies. Digital solutions, such as apps, as well as other traditional suicide prevention interventions could enhance their efforts by supplementing and refining the intervention with personalized support.