nkululeko

Run `npx skills add felixbur/nkululeko --all -g -y` in your terminal to install all skills in this suite globally.

It detects speaker characteristics such as emotion, age, gender, and voice disorders by running machine learning experiments on audio data, all controlled through a simple configuration file.

Yes. You define your dataset, features, and model in a plain INI config file, and Nkululeko handles data loading, training, and evaluation automatically.

It supports sklearn and PyTorch models like SVM, MLP, and fine-tuned transformers, combined with acoustic feature sets such as openSMILE eGeMAPS, Praat, and wav2vec embeddings.

It is built for speech researchers, students, and ML practitioners who want to analyze speaker traits quickly without writing experiment code from scratch.