• Validation of a new device to automatically assess manipulative behaviour in pigs. • Algorithms were developed and validated for the rearing and fattening period. • About 60 % (= sensitivity) of the rope manipulations were detected by an algorithm. • The algorithm reached a specificity and accuracy above 80% for the detection of rope manipulations. • During greater than 65 % of manipulations the manipulating pig was correctly identified. Tail biting is still an important problem in pig husbandry. In addition to addressing the underlying causes, the negative consequences of tail biting can be mitigated by detecting it early e.g. using behavioural changes and implementing intervention measures. Due to the labour intensity of behavioural observations, it would be highly advantageous to detect behavioural changes automatically. This study aimed at developing and validating an automatic device (“Bite-o-Mat”) to assess the individual manipulative behaviour of group-housed pigs. The Bite-o-Mat consisted of a single point load cell (SPLC) that recorded the force with which pigs manipulated a rope, and an UHF-RFID system to individually identify the manipulating pig. Data were recorded in a rearing (8 pigs) and a fattening pen (5 pigs) and validated using twelve hours of video recordings each, distributed across six days. Pigs were observed 596 (rearing) and 277 (fattening) times to manipulate the rope (=manipulation event) in the videos to which the automatically recorded data were compared. Using a linear mixed model, summarising data per manipulation event, it was possible to show that on average stronger forces are exerted on a rope during manipulation events than during times without manipulations and that these forces can be measured by a SPLC (manipulations vs. no manipulations, model estimates in kg; rearing: -0.01 vs. -0.004 (mean), fattening: -0.05 vs. -0.01 (mean)). Based on these results, an algorithm was developed to automatically detect manipulation events using the data of the SPLC. Validation of the algorithm using a second-by-second comparison to the video analysis, showed that it is suitable to detect manipulation events automatically in the rearing and fattening period (sensitivity: 0.60 (rearing and fattening); specificity: 0.87 (rearing), 0.93 (fattening); precision: 0.55 (rearing), 0.58 (fattening); accuracy: 0.81 (rearing), 0.88 (fattening)). In addition, we aimed to identify the manipulating pig with an UHF-RFID antenna. The performance parameters (sensitivity, specificity, accuracy, precision) of the UHF-RFID antenna confirmed that it is sufficiently suitable to detect the pigs within the reading range. Compared with the video recordings, 69% (rearing) and 82% (fattening) of pigs were correctly identified automatically to be the manipulating pigs by the UHF-RFID algorithm. These results indicate that the Bite-o-Mat is not only a promising device to automatically detect manipulative behaviour in group-housed pigs, but also to identify the manipulating animal.