Excessive sighs have been described as one type of dysfunctional breathing (DB). Cardiopulmonary exercise testing (CPET) is one of the diagnostic options for DB and allows for a subjective evaluation of sighs. However, no validated method exists to automatically quantify sighs during CPET. We aimed to develop such a method. We used two Swiss cohorts of patients with persistent dyspnea after SARS-CoV-2 infection using CPET. In the derivation cohort (n = 48), we tested different filters to find the one that was the least influenced by outliers of tidal volume (VT) using a subjective approach. The selected filter (rolling median of 15 values) was applied in the validation cohort (n = 77) to detect spikes of VT above 2 times the value of the associated centered filtered value. Every automatically detected spike of VT from the cohort was analyzed by two experienced raters using continuous volume and flow-over-time graphs reconstructed from high resolution data acquisition. In the validation cohort, 203 automatically detected spikes of VT were visually analyzed by two raters. Of the 203 detected spikes, 199 corresponded to a sigh. The Cohen's Kappa (95% CI) between the raters was 0.89 (0.67-1). In conclusion, we developed a simple automated method for the objective quantification of sighs during CPET. It could be used to establish normative values of sighs during CPET and explore the associations between sighs and symptoms of patients with DB.