Background Viewing self‐harm and suicide‐related images online can precede these behaviours. We reviewed studies of potential impacts and mechanisms associated with viewing self‐harm‐related images on the internet and social media. Method CINAHL, Cochrane Library, EMBASE, HMIC, MEDLINE, PsycArticles, PsycINFO, PubMed, Scopus, Sociological s and Web of Science Core Collection databases were searched for relevant studies from inception to 22 January 2022. Inclusion criteria were English language, peer‐reviewed, empirical studies with data related to impacts of viewing self‐harm images or videos on the internet or social media. Quality and risk of bias were assessed using Critical Appraisal Skills Programme tools. A narrative synthesis approach was used. Results Of the 15 identified studies, all found harmful effects of viewing self‐harm‐related images online. These included escalation of self‐harm, reinforcement of engagement behaviours (e.g. commenting and sharing images), encouragement of social comparison (comparing own self‐harm with others), development of a self‐harm identity, social connection perpetuating or escalating self‐harm, and emotional, cognitive, and physiological impacts triggering self‐harm urges and acts. Nine studies found protective effects, including self‐harm mitigation or reduction, promotion of self‐harm recovery, encouraging social connection and help‐giving, and emotional, cognitive and physiological impacts mitigating or reducing self‐harm urges and acts. Causality of impact was not determined in any study. Most of the studies did not explicitly evaluate or discuss potential mechanisms. Conclusions Viewing self‐harm images online may have both harmful and protective effects, but harmful effects predominated in the studies. Clinically, it is important to assess individual's access to images relating to self‐harm and suicide, and the associated impacts, alongside pre‐existing vulnerabilities and contextual factors. Higher quality longitudinal research with less reliance on retrospective self‐report is needed, as well as studies that test potential mechanisms. We have developed a conceptual model of the impact of viewing self‐harm images online to inform future research.