Hematological profiles are routinely used to assess the health status of animals. Several methods have been developed for blood-cell counting, but typically they are expensive and/or time-consuming. Here, we present a free image-processing software, Mizutama, developed for counting cells in photographs of blood smears. Mizutama uses the thresholding method to transform original photographs into grayscale trinary images. Following a number of parameters, Mizutama searches in the image for cells of a given size, with a nucleus size relative to cytoplasm surface area. The software is not only easy, versatile, and intuitive to handle, but is also fast when counting cells in photographs. Moreover, we show that it is highly accurate, failing to detect only c. 1.4% of avian red cells in ordinary microscopic photographs. The Mizutama application may greatly facilitate the counting of erythrocytes and other blood cells in physiological studies, saving time and money.