camelCase converter — and the five other identifier cases.

Six cases, one phrase

All six conversions share the same first step. The input is broken into words — on spaces, on punctuation, and on the internal capitals of an existing identifier — and then rejoined with a different separator and capitalisation. That shared splitter is what lets you convert between conventions, and not only out of prose.

Note what happened to ID and API: both folded to Id and Api. That is deliberate and it follows the Google Java Style Guide, which spells the procedure out — divide the phrase into words, then “lowercase everything (including acronyms), then uppercase only the first character” of each word. The guide spells its own examples XmlHttpRequest and newCustomerId; the forms it turns away from are XMLHTTPRequest and newCustomerID. Keeping acronyms in full caps produces runs like parseHTTPURLResponse where the eye cannot find the word boundaries.

Every spelling normalises to the same identifier

The splitter treats an existing identifier and a plain English phrase identically, which means converting between conventions needs no special mode. Here are five spellings of one name, each run through the tool:

That is the everyday use: rename a database column to a JavaScript property, turn a class name into a URL slug, or convert a spreadsheet of headings into constants. Paste the whole column and every line is converted independently.

How the splitter handles the awkward cases

The last row is a real limitation, printed here in full. IPv6 is one word that happens to contain an internal capital, and nothing distinguishes that from the genuine word boundary in getUser without a dictionary. The same applies to iOS and eBay. Type them in lowercase — ipv6 — and they come through as a single word.

Which convention belongs where

Identifier casing is not a matter of taste in most ecosystems; it is written down, and linters enforce it. Every row below names the document behind it, and the last column says how far that document goes: spec where a first-party specification, style guide or documentation page states the rule, convention where the practice is near-universal but no first-party document actually mandates it. 3 of the 17 rows carry the weaker label, and the column shows which ones.

Two rows deserve a second look. Go is unusual in having no separate convention for public and private names: the initial capital is the visibility marker, so UserName is exported and userName is not, and underscores are ruled out entirely. URLs are kebab-case for a documented reason and not a stylistic one: Google's guidance is to use hyphens rather than underscores in URLs, because a hyphen reads as a word separator and an underscore does not.

Practical notes

One identifier per line. The converter treats each line separately, so a paragraph of prose becomes one very long identifier — which is almost never what you want. Split your input into lines first and it does the right thing for every one of them.

Accents and non-Latin letters are kept. The splitter works on Unicode letters, not ASCII, so día laboral becomes díaLaboral. Most languages accept that in an identifier; if your target does not, strip the accents before converting.

Undo is per conversion. Try camelCase, decide you wanted kebab-case, and just press the other button — the splitter re-derives the same words either way. Undo is there for when you want the original phrase back, punctuation and all.