Background: Peer support for mental health is recommended across international policy guidance and provision and has recently been expanded in countries including the UK. We conducted a systematic umbrella review, summarising the evidence from published reviews of the: 1) effectiveness, 2) implementation, and 3) experiences of paid peer support approaches for mental health. Methods: We searched MEDLINE, EMBASE, PsycINFO, The Campbell Collaboration, and The Cochrane Database of Systematic Reviews (from January 2012-November 2022) for published reviews of paid peer support interventions for mental health. Review quality was assessed using the AMSTAR2. Results were synthesised narratively, with implementation reported in accordance with the Consolidated Framework for Implementation Research (CFIR). Results: We included 35 reviews: systematic reviews with (n=13) or without (n=13) meta-analysis, systematic reviews with qualitative synthesis (n=3), and scoping reviews (n=6). The reviews included 426 primary studies and between 95-40,927 participants. Most reviews were low or critically low (97%) quality, one review was high quality. Effectiveness was investigated in 23 reviews. While results were mixed, there was some evidence from meta-analyses that peer support may improve depression symptoms (particularly in perinatal depression), self-efficacy, and recovery. Factors promoting successful implementation, investigated in 9 reviews, included adequate training and supervision, a recovery-oriented workplace, strong leadership, and a supportive and trusting workplace culture with effective collaboration. Barriers to implementation included lack of time, resources and funding, and a lack of recognised PSW certification. Experiences of peer support were explored in 11 reviews, with 3 overarching themes: i) what the peer support role could contribute, including recovery and improved wellbeing for both service users and peer support workers (PSWs); ii) confusion over the PSW role, including role ambiguity and unclear boundaries; and iii) organisational challenges, including low pay for PSWs, negative non-peer staff attitudes, and lack of support and training. Discussion: Peer support may be effective at improving some clinical outcomes, self-efficacy, and recovery outcomes for some people. Evidence suggests that certain populations, such as perinatal populations, may especially benefit from peer support. Potential strategies to successfully implement PSWs in healthcare settings include coproduction, with clearly defined PSW roles, a receptive hierarchical structure and staff, strong leadership, and appropriate PSW and staff training with clinical and/or peer supervision alongside safeguarding. Services could also benefit from clear, coproduced, setting specific implementation guidelines for PSW. PSW roles tend to be poorly defined and associations between content of PSW interventions and their impacts needs further investigation. Future research in this area should reflect the priorities of those directly involved in peer support, either as providers or service users.