BACKGROUND An often-used tool to measure adherence to antiretroviral therapy (ART) is the Medication Event Monitoring System (MEMS), an electronic pill-cap that registers date and time of pill-bottle openings. Despite its strengths, MEMS-data can be compromised by inaccurate use and acceptability problems due to its design. These barriers remain, however, to be investigated in resource-limited settings. We evaluated the feasibility and acceptability of using MEMS-caps to monitor adherence among HIV-infected patients attending a rural clinic in Tanzania's Kilimanjaro Region. METHODS Eligible patients were approached and asked to use the MEMS-caps for three consecutive months. Thereafter, qualitative, in-depth interviews about the use of MEMS were conducted with the patients. MEMS-data were used to corroborate the interview results. RESULTS Twenty-three of the 24 patients approached agreed to participate. Apart from MEMS-use on travel occasions, patients reported no barriers regarding MEMS-use. Unexpectedly, the MEMS-bottle design reduced the patients' fear for HIV-status disclosure. Patients indicated that having their behavior monitored motivated them to adhere better. MEMS-data showed that most patients had high levels of adherence and there were no bottle-openings that could not be accounted for by medication intake. Non-adherence in the days prior to clinic visits was common and due to the clinic dispensing too few pills. CONCLUSIONS MEMS-bottle use was readily accepted by patients. Although the MEMS-bottle was used accurately by most patients, patients need to be more explicitly instructed to continue MEMS-use when travelling. Even HIV-clinics with sufficient staff and free medication may impose structural adherence barriers by supplying an insufficient amount of pills.