Loneliness and social isolation are currently among the most challenging social issues. Given their detrimental impact on physical and mental health, identifying feasible and sustainable interventions to alleviate them is highly important. Friendly visiting, a befriending intervention whereby older persons are matched with someone who visits them on a regular basis, seems promising. However, it is unclear if face-to-face (F2F) friendly visiting by a volunteer (FVV) is effective at reducing loneliness or social isolation, or both. To assess the effect of F2F FVV on feelings of loneliness, social isolation (primary outcomes) and wellbeing (i.e., life satisfaction, depressive symptom experiencing and mental health; secondary outcomes) in older adults. We searched six electronic databases up until 11 August 2021. We also consulted 15 other resources, including grey literature sources and websites of organizations devoted to loneliness and ageing, between 25 October and 29 November 2021. We included experimental and observational studies that quantitatively measured the effect of F2F FVV, compared to no friendly visiting, on at least one of following outcomes in older adults (≥60 years of age): loneliness, social isolation or wellbeing. Two reviewers independently performed study selection, data extraction and synthesis, risk of bias and GRADE assessment. If outcomes were measured multiple times, we extracted data for one short-term (≤1 month after the intervention had ended), one intermediate-term (>1 and ≤6 months), and one long-term time point (>6 months). Data from randomized controlled trials (RCTs) and non-RCTs were presented and synthesized separately. Synthesis was done using vote counting based on the direction of effect. Nine RCTs and four non-RCTs, conducted primarily in the United States and involving a total of 470 older adults (mean or median ages: 72-83 years), were included. All studies were limited in size (20-88 participants each). Programmes lasted 6-12 weeks and mostly involved weekly visits by undergraduate students to community-dwelling older adults. Visits consisted mainly of casual conversation, but sometimes involved gameplaying and TV-watching. All studies had major shortcomings in design and execution. The current evidence about the effect of F2F FVV on loneliness in older adults is very uncertain, both in the short (one RCT in 88, and one non-RCT in 35 participants) and intermediate term (one RCT in 86 participants) (both very low-certainty evidence). The same goes for the effects on social isolation, again both in the short (one RCT in 88, and two non-RCTs in 46 participants) and intermediate term (two non-RCTs in 99 participants) (both very low-certainty evidence). Similarly, there is a lot of uncertainty about the effect of F2F FVV on outcomes related to wellbeing (all very low-certainty evidence). Due to the very low-certainty evidence, we are unsure about the effectiveness of F2F FVV with regard to improving loneliness, social isolation, or wellbeing in older adults. Decision-makers considering implementing FVV should take into account this uncertainty. More and larger high-quality studies that are better designed and executed, and preferably conducted in various settings, are needed.