Sepsis requires rapid identification of the infectious microorganism so that targeted antimicrobial treatment can be started as early as possible. Conventional blood culture remains important, but it may require one or more days before preliminary identification is available. This study evaluated manual and semi-automated QIAcube DNA extraction workflows for recovering microbial DNA from Luria Broth-grown bacterial cultures and from human blood samples spiked with Escherichia coli. The extracted DNA was assessed using NanoDrop and Qubit measurements, and selected samples were cleaned with AMPure beads before Oxford Nanopore MinION library preparation and sequencing. Extraction performance was evaluated using DNA concentration, A260/A280 and A260/A230 purity ratios, hands-on time, turnaround time, sequencing output, and Kraken 2 taxonomic classification in EPI2ME. Manual extraction produced higher DNA concentrations for several Escherichia coli culture samples than the semi-automated QIAcube workflow, but the results were more variable between replicates. The QIAcube workflow reduced hands-on time and produced more standardized handling, although total turnaround time was longer because of instrument processing. Blood-derived samples were more challenging because bacterial DNA was present at low abundance together with host-derived human DNA. Nanopore sequencing generated sufficient read yield and quality for taxonomic analysis, and expected bacterial taxa, including Escherichia coli and Staphylococcus epidermidis, were detected in culture-derived and selected blood-derived samples. Classification at a Kraken confidence threshold of 0 improved species-level detection compared with 0.1, although unclassified and Homo sapiens reads remained. Overall, the study supports the potential of nanopore sequencing for rapid microbial identification while showing that host-DNA depletion, extraction efficiency, and low-biomass sample handling require further optimization before clinical implementation.