Over the last decade there has been an explosion of interest in mindfulness-based self-help (MBSH) interventions. While widely available and extensively promoted, there is little consensus on their impact in public health or healthcare contexts. We present a systematic review and meta-analyses of 83 randomized controlled trials, comparing unguided MBSH to control conditions on outcomes of depression, mindfulness, anxiety, stress and/or wellbeing/ quality of life. A random effects model was used to compute post-intervention, between-groups effect sizes for each outcome. MBSH demonstrated small, statistically significant effects at post-interventions for outcomes of depression (g = -0.23), mindfulness (g = 0.37) anxiety (g = -0.25), stress (g = -0.41) and wellbeing/ quality of life (g = 0.34). Significant effects were retained at follow-up for mindfulness, stress and wellbeing/ quality of life but not for depression or anxiety. Planned moderator analyses demonstrated significantly larger effects of MBSH when compared to inactive, versus active-control conditions on all outcomes except wellbeing/ quality of life, and non-digital MBSH interventions demonstrated significantly greater effects on depression, mindfulness and wellbeing/ quality of life outcomes than digitally-delivered MBSH. When studies that utilised samples selected for mental and physical health-related difficulties were respectively compared to studies that utilised unselected samples, no significant moderation effects were observed. In sum, these findings provide evidence for the effectiveness of unguided MBSH in public health settings and the practical, access-related implications of this are discussed.