Abstract The Amazon Basin is the largest in the world and for many of its inhabitants fishing is a crucially important source of food and income. However, the benefits derived from mostly informal, part time, seasonal, and subsistence-based activities such as fishing are largely invisible to policy-makers, and addressing the threats to aquatic habitats, ecosystem functioning, and fisheries is frequently given low priority in national development agendas. To estimate the total extraction of fish in the Amazon Basin, we reviewed various publications and databases with quantitative data related to the landings, trade, and consumption of fish. We estimated the total landings to be between 422,000 and 473,000 t yr-1 in live weight, near to previous estimates. Almost 75%, however, represented landings in the Brazilian part of the basin, and there is a very large margin of uncertainty given that, among other issues, there are no recent official statistics on commercial landings. Conversely, landings in Bolivia, Colombia, and Ecuador seem to be considerably higher than previously thought. In all Amazonian countries, from 50% up to almost 100% of the fishing takes place for personal consumption, implying that it is not recorded in official landing statistics. Available time series data indicate that fish consumption in Brazil, as well as commercial fish landings in Peru and Colombia, have declined in recent years. Local case studies, national household consumption and expenditure surveys (HCES), and market surveys, all have their benefits and drawbacks. Some minor modifications of HCES procedures could considerably improve their usefulness for estimating fish landings in the Amazon Basin. To improve such estimates, we propose that local case studies be used to calibrate HCES data, thus combining the high precision of local case studies with the wide coverage and representativity of HCES data