HIV self-testing (HIVST) addresses barriers to HIV diagnosis among men, but current approaches to distributing HIVST kits only reach a subset of the men requiring testing.We conducted a pilot trial of the secondary distribution of HIVST kits through peer networks in fishing communities of Buliisa district (Uganda). We recruited distributors ("seeds") among male patients of a health facility, and among community members. Seeds were trained in HIVST and asked to distribute up to five kits to their peers ("recruits"). Recruits were referred to the study using a coupon, and asked to return the HIVST kit (used or unused). The accuracy of HIVST was measured against a confirmatory test conducted by a health worker. We conducted audio computer assisted self-interviews to measure the occurrence of adverse events, and evaluate the potential yield of peer-delivered HIVST. We also assessed how seeds and recruits rated their experience with peer-distributed HIVST.Nineteen seeds offered an HIVST kit to 116 men, and 95 (81.9%) accepted the offer. No recruit reported coercion, but two seeds experienced hostility from recruits or their family members. The sensitivity of peer-distributed HIVST, as interpreted by recruits, was 100%, and its specificity was 92.8%. Among recruits, 29 had never tested (25.8%), and 42 (44.2%) had tested more than a year ago. Three men living with HIV learned their status through peer-distributed HIVST (yield = 1 new diagnosis per 6.3 seeds). Most recruits (85/88) and seeds (19/19) reported that they would recommend HIVST to their friends and family. All seeds stated that they would accept acting as peer distributors again.This novel peer-based distribution model of HIVST is safe, and has high uptake. It could help reduce the gender gap in HIV testing in under-served fishing communities in Uganda and elsewhere.