Field guides
Repair copied text without guessing
Each guide documents a reproduced failure, the rule TextClean applies, what the rule cannot know, and the checks a person should still make.
Clean the damage, not the document
Choose a cleanup method from the source, preserve structure, and use a three-pass review.
PDF text repairRepair PDF line breaks without flattening the document
A tested method for joining wrapped PDF lines while preserving paragraphs, common headings, lists, and deliberate hyphens.
Unicode inspectionInvisible Unicode characters: inspect before removing
Identify zero-width spaces, joiners, direction controls, variation selectors, tags, and special spaces without damaging multilingual text.
Formatting cleanupClean AI formatting without rewriting the text
Remove Markdown, citation markers, and selected punctuation while keeping cleanup separate from authorship and rewriting.
HTML extractionConvert HTML to readable text without keeping script content
Understand which HTML elements should create paragraph breaks, which content should be discarded, and why extraction is not sanitization.
Destination formattingFormat cleaned text for email, messages, blogs, and documents
Preserve headings, lists, quotations, code, and paragraph spacing while preparing the same cleaned text for different destinations.