Dependent upon the statistical significance p-value and statistical power, the sample size estimation is widely used in various experimental sciences. Nevertheless, the p-value based paradigm, which has resulted in numerous fake conclusions that originate partly from insufficient sample sizes, has been widely criticized in recent years for serious problems. Therefore, I developed a platform-independent computational tool, SampSizeCal, for sample sizes in the paradigm of new statistics. In this tool, both default p-values and the maximum p-values were greatly enhanced, which will lead to the reasonable increase of sample sizes. The computational tool harbors more than 120 sample size methods for experimental designs. SampSizeCal includes both online and offline versions, and can be used for various computing devices (PCs, iPads, smartphones, etc.), operating systems (Windows, Mac, Android, Harmony, etc.) and web browsers (Chrome, Firefox, Sougo, 360, etc). It is currently the most comprehensive platform-independent computational tool for sample sizes, and can be used in experimental sciences such as medicine (clinical medicine, experimental zoology, public health, pharmacy, etc.), biology, ecology, agronomy, psychology and engineering technology.