Testing for SARS-CoV-2 is one of the most important assets in COVID-19 management and mitigation. At the onset of the pandemic, SARS-CoV-2 testing was uniquely performed in central laboratories using RT-qPCR. RT-qPCR relies on trained personnel operating complex instrumentation, while time-to-result can be lengthy (e.g., 24 to 72 h). Now, two years into the pandemic, with the surge in cases driven by the highly transmissible Omicron variant, COVID-19 testing capabilities have been stretched to their limit worldwide. Rapid antigen tests are playing an increasingly important role in quelling outbreaks by expanding testing capacity outside the realm of clinical laboratories. These tests can be deployed in settings where repeat and rapid testing is essential, but they often come at the expense of limited accuracy and sensitivity. Reverse transcription loop-mediated isothermal amplification (RT-LAMP) provides a number of advantages to SARS-CoV-2 testing in standard laboratories and at the point-of-need. In contrast to RT-qPCR, RT-LAMP is performed at a constant temperature, which circumvents the need for thermal cycling and translates into a shorter analysis time (e.g., <1 h). In addition, RT-LAMP is compatible with colorimetric detection, facilitating visualization and read-out. However, even with these benefits, RT-LAMP is not yet clinically deployed at its full capacity. Lack of automation and integration of sample preparation, such as RNA extraction, limits the sensitivity and specificity of the method. Furthermore, the need for cold storage of reagents complicates its use at the point of need. The developments presented in this work address these limitations: We describe a fully automated SARS-CoV-2 detection method using RT-LAMP, which also includes up-front lysis and extraction of viral RNA, performed on a centrifugal platform with active pneumatic pumping, a disposable, all-polymer-based microfluidic cartridge and lyophilized reagents. We demonstrate that the limit of detection of the RT-LAMP assay itself is 0.2 copies per μL using N and E genes as target sequences. When combined with integrated RNA extraction, the assay sensitivity is 0.5 copies per μL, which is highly competitive to RT-qPCR. We tested the automated assay using 12 clinical swab specimens from patients and were able to distinguish positive and negative samples for SARS-CoV-2 within 60 min, thereby obtaining 100% agreement with RT-qPCR results.