CorpusPrep Keeps a Paper Trail for Text Cleaning

Field notes

CorpusPrep reads source files in place, segments them for review, and waits for an explicit cleaning request before producing output. Its presets range from retaining everything to removing publisher apparatus or headings, while a custom mode lets users select sections individually. The interface supports TXT, Markdown, DOCX, EPUB, HTML, and PDF inputs, and presents synchronized source and cleaned panes with removed passages marked. A preprocessing log records the reason for each removal.