Background Robots are seen as one of the possibilities to provide care to older people, especially when care by humans becomes scarce and too expensive. In the past sixty years, however, robots have diversified into many different types with a very broad range of functionalities. Despite this large variety of operational robots that have been developed in the past decades, the majority of existing types of robot will currently not be capable of carrying out reliably and effectively meaningful care tasks. For one thing, the communication between the more traditional types of robots and humans is severely underdeveloped. This is only logical as robots were intended to replace human practice in which human-robot communication was either undesirable or of limited value. Aim Fortunately, exceptions to this are increasingly appearing, and these will be treated. At this stage it will not be surprising that an overview of types of robots for care will be a categorization of types of robots, rather than types of care. As it is, current prototypes of robots for care also have different functionalities, and with those, also a very different architecture indicating that no single planned robot can perform all care tasks that are necessary or desired. This manuscript presents an overview of the main types of care robots that are currently under development, and partly even operative in actual care locations. Results It is evident that also the care robots in actual use still have quite limited capabilities, and, coupled with this, a short endurance of about an hour. Yet it also has to be concluded that even the simplest types of robots with a very limited behavioural repertoire can be successful with cases of cognitive decline and early dementia. In view of the potential capabilities of robots a categorization of care activities is proposed that can be stated as (i) physical assistance, (ii) social assistance and (iii) medical assistance. Physical assistance can be provided by relatively large and heavy-weight robots that can lift and transport people with limited mobility. For the great majority of care robots, social assistance is envisaged, to be realized by humanoid robots, looking like real humans but mostly half-size or less. Active research in this field is focusing on robot locomotion, gestures, speech recognition, understanding and production, face recognition and autonomy. Somewhat disconcerting is the observation that medical assistance robots are practically absent in current development efforts, though it can be argued that there are definitely many relatively simple and frequently occurring medical actions that could be successfully carried out by a robot.