Metaprop is a Stata command for performing meta-analysis of binomial data. It extends the existing metan procedure by providing methods specific to binomial data, including exact binomial and score test-based confidence intervals. It addresses issues with normal approximation methods by using the binomial distribution to model within-study variability or applying the Freeman-Tukey double arcsine transformation to stabilize variances. Metaprop was applied to two published meta-analyses: one on HPV infection prevalence in women with ASC-US and another on cure rates after cervical precancer treatment with cold coagulation. The first meta-analysis found a pooled HPV prevalence of 43% (95% CI: 38%-48%), while the second found a pooled cure rate of 94% (95% CI: 86%-97%). Metaprop allows inclusion of studies with 0% or 100% proportions, ensuring confidence intervals remain within admissible values. It uses the logistic-normal random-effects model, which is recommended for binomial data. Metaprop provides exact confidence intervals and score confidence intervals, which are more reliable than Wald intervals, especially for proportions near the boundaries. The program is available for download and can be installed in Stata. It is designed to handle binomial data and provides accurate results for meta-analysis of proportions.Metaprop is a Stata command for performing meta-analysis of binomial data. It extends the existing metan procedure by providing methods specific to binomial data, including exact binomial and score test-based confidence intervals. It addresses issues with normal approximation methods by using the binomial distribution to model within-study variability or applying the Freeman-Tukey double arcsine transformation to stabilize variances. Metaprop was applied to two published meta-analyses: one on HPV infection prevalence in women with ASC-US and another on cure rates after cervical precancer treatment with cold coagulation. The first meta-analysis found a pooled HPV prevalence of 43% (95% CI: 38%-48%), while the second found a pooled cure rate of 94% (95% CI: 86%-97%). Metaprop allows inclusion of studies with 0% or 100% proportions, ensuring confidence intervals remain within admissible values. It uses the logistic-normal random-effects model, which is recommended for binomial data. Metaprop provides exact confidence intervals and score confidence intervals, which are more reliable than Wald intervals, especially for proportions near the boundaries. The program is available for download and can be installed in Stata. It is designed to handle binomial data and provides accurate results for meta-analysis of proportions.