The diagnosis of COVID-19 is based on detection of SARS-CoV-2 in oro-/nasopharyngel swabs, but due to discomfort and minor risk during the swab procedure, detection of SARS-CoV-2 has been investigated in other biological matrixes. In this proof-of-concept study, individuals with confirmed SARS-CoV-2 infection performed a daily air sample for five days. Air samples were obtained through a non-invasive electrostatic air sampler. Detection of SARS-CoV-2 RNA was determined with qRT-PCR. The association of positive samples with different exposures was evaluated through mixed-effect models. We obtained 665 air samples from 111 included participants with confirmed SARS-CoV-2 infection. Overall, 52 individuals (46.8%) had at least one positive air sample, and 129 (19.4%) air samples were positive for SARS-CoV-2. Participants with symptoms or a symptom duration ≤ four days had significantly higher odds of having a positive air sample. Cycle threshold values were significantly lower in samples obtained ≤ 4 days from symptom onset. Neither variant of SARS-CoV-2 nor method of air sampling were associated with a positive air sample. We demonstrate that SARS-CoV-2 is detectable in human breath by electrostatic air sampling with the highest detection rate closest to symptom onset. We suggest further evaluation of the air sampling technique to increase sensitivity.