Background Antibody-based serological tests which target households and communities can estimate the true extent of infection in a population. It minimizes the biases of facility-based selective testing and generates scientific data on disease transmission through household asymptomatic cases. The objective of this study was to determine the seroprevalence and trend of SARS-CoV-2 in a densely populated urban community of Karachi. Methods Three serial cross-sectional surveys were conducted in November 2020, February 2021, and December 2021 in Karachi’s District East. Households were selected to provide serum samples for Elecsys® immunoassay for the detection of SARS-CoV-2 antibodies. All household members were eligible to participate regardless of age and infection status. Bayesian regression was used to adjust for assay performance and estimate seroprevalence. Results We enrolled 1506 participants from 501 households. In November 2020, adjusted seroprevalence was estimated as 24.0% (95% confidence interval, CI=18.0-31.0), compared to 53.9% (95% CI=45.5-63.2) in February. In December 2021, it increased to 84.9% (95% CI=78.5-92.3). The conditional risk of infection was 41% (95% CI=29.9-51.6), 56.7% (95% CI=50.4–62.6) and 77.8% (95% CI=73.0-81.7) in surveys 4, 5, and 6 respectively. Only 18.7% of participants who had reactive antibodies for COVID-19 were symptomatic. Conclusions An increase in seroprevalence estimates in Karachi’s District East was observed over time. Community-based seroprevalence studies help to estimate the true proportion of the population that has been infected and predicts the spread of the disease in similar settings.