Text Case Converter

The Text Case Converter transforms input across 17 case styles split into three groups in the UI: Basic (lowercase, UPPERCASE, Title Case, Sentence case, Capitalize Words), Programming (camelCase, PascalCase, snake_case, kebab-case, CONSTANT_CASE, dot.case, path/case, Header-Case) and Special (url-slug, iNVERT cASE, aLtErNaTiNg, RaNdOm CaSe). The hard part of any case converter is word-boundary detection, and this one uses a four-step tokenizer (called getWords in the code): it inserts a space between a lowercase-then-uppercase pair (so 'helloWorld' splits into 'hello World'), inserts a space at an acronym-to-word boundary using the pattern that matches a run of capitals followed by a capital+lowercase (so 'XMLParser' splits as 'XML Parser'), replaces any of the delimiters - _ . / with spaces, and finally splits on whitespace. All the programming styles (snake_case, kebab-case, dot.case, path/case, CONSTANT_CASE, camelCase, PascalCase, Header-Case) are built by re-joining those detected words with the appropriate separator.

Because the tokenizer lowercases each word before re-joining for most programming styles (word.toLowerCase() for snake/kebab/dot/path, word.toUpperCase() for CONSTANT, and char-0-uppercase for camel/Pascal/Header), conversions are deliberately destructive to original capitalization: an input acronym like HTTP survives the split as its own word but is flattened to 'http' in snake_case or 'Http' in PascalCase. This makes round-trips lossy - going to Title Case and back to camelCase will not restore an original ALL-CAPS acronym. The boundary rules are also where ambiguity shows up: 'IOError' is split via the acronym rule into 'IO Error', but a single-letter prefix or trailing digits follow the literal regex, not English intuition. Title Case is the one style that does NOT use the word tokenizer - it splits on plain spaces, lowercases everything, then re-capitalizes, while skipping a hardcoded list of 18 minor words (a, an, and, as, at, but, by, for, in, nor, of, on, or, so, the, to, up, yet) unless they are the first word. Sentence case capitalizes only the first letter and any letter following . ! or ?, and url-slug lowercases, strips every character outside word/space/hyphen, collapses runs of spaces and hyphens, and trims leading/trailing hyphens.

The two random styles behave differently and it matters: aLtErNaTiNg is deterministic by character index (even positions lowercase, odd positions uppercase via i % 2), while RaNdOm CaSe calls Math.random() per character, so it produces a different result on every keystroke and is not reproducible. The tool runs entirely in the browser with no network calls. A Trim checkbox strips leading/trailing whitespace before conversion; a file picker accepts .txt, .md, .json and .csv and loads the contents into the input via FileReader; each result can be copied or downloaded as a converted-{caseType}-{timestamp}.txt plain-text Blob. Every copy action also pushes an entry into an in-memory history that keeps the last 10 conversions with a Restore button, and the input plus view settings are encoded in the URL via the shareable-config button, so a specific input and selected case can be linked directly. A live counter reports character, word and line counts on the input.