This paper introduces Protosody, a semi-automated protocol developed using Python and Praat Scripting Language. This interdisciplinary approach combines computational methods with (foreign) language speech studies to enhance the extraction of prosodic-acoustic features in sound-transcription studies. The protocol operates on pairs of ‘.wav/.flac’ audio files and ‘.vtt’ files, transforming the linguistic data, numbers, punctuation, and orthographic diacritics into plaintext and retagging the files. These files are subsequently uploaded to a phonetic forced aligner. The protocol processes the returned forced-aligned TextGrids, converting phonemic-sized units into syllabic or higher speech units such as sentences or utterances, and defining the tonal range for each higher speech unit. The method extracts a comprehensive set of 74 features, encompassing four factors (Speaker_ID, Language-Dialect-Accent, Sex, Sentence-Utterance), 30 rhythm metrics, and 40 prosodic-acoustic features.