Evidence for acute caffeine supplementation in swimming remains inconsistent across reviews, likely due to between-study heterogeneity and methodological differences. Five databases were searched for randomized placebo-controlled crossover trials exploring the effects of caffeine on swimming performance in competitive swimmers. Performance outcomes were pooled as standardized mean differences (SMD; Hedges' g), with time outcomes sign-reversed so that positive values indicated improved performance. Dependence among multiple outcomes within studies was addressed using three-level random-effects models (REML). As a supplementary, interpretable metric, log ratios of means (lnRoM) were meta-analysed and expressed as percentage change. Post-exercise blood lactate (mmol/L) was pooled as mean differences (MD). Risk of bias was assessed using RoB 2 and certainty of evidence using GRADE. Prespecified exploratory moderator analyses examined potential effect modification (dose, timing, distance, administration form, stroke, gender, and athlete level). Thirteen studies were included (144 men and 48 women), which contributed 28 performance effect sizes. Caffeine improved swimming performance (SMD = 0.57, 95% CI: 0.20 to 0.94; p = 0.005), corresponding to a +1.71% improvement in lnRoM (95% CI: +1.01% to +2.41%; p