Background: Metabarcoding environmental DNA with high-throughput sequencing is a state-of-the-art method to assess biodiversity and to uncover dark taxa. MinION is the first handheld sequencer that can be taken into the field for on-site metabarcoding. This research aims to answer if bioinformatics can solve the issues that arise because of the higher error rate of MinION data. Results: Biodiverse samples with a presumed large portion of dark taxa were selected from the Dutch Caribbean. The cytochrome oxidase 1 gene (CO1) is used as a barcode for identification at the species level or higher levels. Generating a consensus sequence from closely related sequences resulted in minimized random errors and increased species identification of 175% compared to unclustered MinION data. Additional to the formulation of the workflow, an ecological analysis was conducted that revealed co-occurrence of species in similar habitats, and that the proportion of dark taxa in the sampled region is 81.87%. Conclusion: Although the workflow did not attain the results that the existing Illumina workflows do, the potential is evident. The high proportion of dark taxa in the sampled region of Statia and the Saba Bank indicates the need for continued barcoding of species in the Dutch Caribbean to resolve database limitations.