BACKGROUND: Although several types of in vitro digestibility assays have been applied to nutritional evaluation of feeds for aquatic organisms, all of them are based on the use of closed reactors and do not simulate the gastric phase of the digestion. Our objective was to evaluate the suitability of a gastrointestinal model based on the use of a digestion cell provided by a semi-permeable reaction chamber, which allows continuous removal of digestion products as they are produced. We tested the effects of some factors, like the inclusion of a gastric phase, reaction temperature or bile salts on the hydrolysis of feed proteins by fish enzymes.