OBJECTIVE: To assess whether coronavirus disease 2019 (COVID-19) is associated with menstrual cycle length changes and, if so, how that compares with those undergoing vaccination or no event (control). METHODS: We conducted a retrospective cohort analysis in which we analyzed prospectively tracked cycle-length data from users of a period tracker application who also responded to a survey regarding COVID-19 symptoms and vaccination. We restricted our sample to users aged 16–45 years, with normal cycle lengths (24–38 days) and regular tracking behavior during the five cycles around COVID-19 symptoms or vaccination or a similar time period for those experiencing no event (control group). We calculated the within-user change in cycle length (days) from the three consecutive cycles preevent average (either vaccination, disease, or neither; cycles 1–3) to the event (cycle 4) and postevent (cycle 5) cycles. We used mixed-effects models to estimate the age- and country-adjusted difference in change in cycle length across the groups. RESULTS: We included 6,514 users from 110 countries representing 32,570 cycles (COVID-19 symptoms: 1,450; COVID-19 vaccination: 4,643; control: 421). The COVID-19 cohort experienced a 1.45-day adjusted increase in cycle length during cycle 4 (COVID-19) compared with their three preevent cycles (95% CI 0.86–2.04). The vaccinated group experienced a 1.14-day adjusted increase in cycle length during cycle 4 (COVID-19 vaccine) compared with their preevent average (95% CI 0.60–1.69). The control group (neither vaccine nor disease) experienced a 0.68-day decrease (95% CI −1.18 to −0.19) in a similar time period. Post hoc tests showed no significant differences in the magnitude of changes between the COVID-19 and vaccination cohorts. In both cohorts, cycle length changes disappeared in the postevent cycle. CONCLUSION: Experiencing COVID-19 is associated with a small change in cycle length similar to COVID-19 vaccination. These changes resolve quickly within the next cycle.