Arsenic in groundwater is recognised as a threat to public health worldwide and specifically in rural areas of several developing countries (e.g. Bangladesh). Different treatment technologies are available or under investigation that, when properly operated, can reduce arsenic concentrations to a sufficiently low level. These technologies are inappropriate for rural communities in developing countries where arsenic-related problems are most pronounced, and where centralised water treatment facilities are frequently absent. Arsenic-related health hazards are associated with the use of arsenic-contaminated water for drinking and food preparation, corresponding to only 2–4 litres per capita per day. Arsenic removal in rural areas, where centralised water treatment and distribution facilities are unavailable, should consequently be carried out at household level and the system applied should be simple and affordable. Arsenic adsorption capacities of different virgin and iron-coated materials were first examined in batch adsorption experiments. Both iron-coated sand (ICS) and iron-impregnated granular activated carbon (IGAC) show high As(III) and As(V) adsorption capacities that decrease at high pH values. Three simple point-of-use systems, namely, a ‘tea bag’, a ‘coffee filter’ and a simple ‘family filter’, all with iron-coated sand or iron-impregnated granular activated carbon were tested. The ‘tea bag’ system was not successful because too long an adsorption time was required. The ‘coffee filter’ performed better, however, results obtained with the simple ‘family filter’ were most promising. The simple ‘family filter’ containing 2 litres of IGAC is expected to supply a family in rural Bangladesh with arsenic-free water for drinking and cooking for, on average, 18 months without adsorbent replacement.