Clinical researchers make use of experimental models for mental disorders. In many cases, these models use stimuli that are relevant to the disorder under scrutiny, which allows one to experimentally investigate the factors that contribute to the development of the disorder. For example, one might use spiders or spider-like stimuli in the study of specific phobia. More broadly, researchers often make use of real-world stimuli such as images of animals, geometrical shapes or emotional words. However, these stimuli are often limited in their experimental controllability and their applicability to the disorder in question. We present a novel set of animal-like stimuli, called Fribbles, for use within behavioural research. Fribbles have desirable properties for use in research because they are similar to real-world stimuli, but due to their novelty, participants will not have had previous experience with them. They also have known properties that can be experimentally manipulated. We present an investigation into similarity between Fribbles in order to illustrate their utility in research that relies on comparisons between similar stimuli. Fribbles offer both experimental control and generalisability to the real world, although some consideration must be made concerning the properties that influence similarity between Fribbles when selecting them along a dimension of similarity.