We sequenced the standard DNA barcode gene fragment in 694 newly collected specimens, representing 243 species level Operational Barcode Units (OBUs) of freshwater fishes from Bangladesh. We produced coi sequences for 149 out of the 237 species already recorded from Bangladesh. Another 83 species sequenced were not previously recorded for the country, and include about 30 undescribed or potentially undescribed species. Several of the taxa that we could not sample represent erroneous records for the country, or sporadic occurrences. Species identifications were classified at confidence levels 1(best) to 3 (worst). We propose the new term Operational Barcode Unit (OBU) to simplify references to would-be DNA barcode sequences and sequence clusters. We found one case where there were two mitochondrial lineages present in the same species, several cases of cryptic species, one case of introgression, one species yielding a pseudogene to standard barcoding primers, and several cases of taxonomic uncertainty and need for taxonomic revision. Large scale national level DNA barcode prospecting in high diversity regions may suffer from lack of taxonomic expertise that cripples the result. Consequently, DNA barcoding should be performed in the context of taxonomic revision, and have a defined, competent end-user.