A binomial distribution is given by the two parameters n and p. Often n is known so that only p has to be estimated. If success has been observed k times the maximum likelihood estimate of p is just k/n. For this kind of estimation specialized software is not needed.