Soft biometric traits, e.g. gender or eye colour, can be effectively used to assist traditional biometric recognition systems. The facial region is a particularly rich source of such features. The goal of this project was to implement an API framework for extraction of soft biometric features from facial images. Many similar commercial tools exist, but often cannot be used due to licensing constraints with biometric research databases. The framework developed in this project is locally-run, flexible and easily extensible, aggregrates self-developed soft-biometrics detection functions and integrates already available methods from various open-source libraries.