Food mislabelling is a growing world-wide problem that is increasingly addressed through the authentication of ingredients via techniques like mass spectrometry or DNA-sequencing. However, traditional DNA sequencing methods are slow, expensive, and require well-equipped laboratories. We here test whether these problems can be overcome through the use of Nanopore sequencing. We sequenced 92 single and 13 mixed-species samples bought in supermarkets and restaurants in Singapore which has a large and diverse seafood trade. We successfully obtained DNA barcodes for 94% and 100% of the single- and mixed-species products after correcting the numerous sequencing errors of MinION reads with a correction pipeline optimized for DNA barcodes. We find comparatively low levels of clear-cut mislabelling for single-species samples (7.6 %) while the rates are higher for mixed-species samples (38.5 %). These low rates are somewhat deceptive, however, because of the widespread use of vague common species names that do not allow for a precise assessment of the expected ingredients. With regard to the clearly mislabelled single-species products, higher-value products (e.g., prawn roe, wild-caught Atlantic salmon, halibut) are replaced with lower-value ingredients (e.g., fish roe, Pacific salmon, arrowtooth flounder) while more serious problems are observed for mixed-species samples. Cuttlefish and prawn balls repeatedly contained pig DNA and 100% of all mixed samples labelled as containing crustaceans (‘crab’, ‘prawn’, ‘lobster’) only yielded fish barcodes. We conclude that there is a need for more regular testing of seafood samples and suggest that due to speed and low-cost, MinION would be a good instrument for this purpose. We also emphasize the need for developing clearer labelling guidelines.