Browser-based · Case conversion and cleanup
Change the case. Keep the meaning.
Convert, clean and standardise text for writing, publishing and code. Motifuse implements four title-case style guides separately, protects acronyms and brand names, and marks every word it changed.
Case conversion
Convert, clean and compare
Paste text, pick a format, and see exactly which words changed before you copy it back. Counting the result instead? Word Counter carries the metrics side of the same document.
Source text
Words
0
Characters
0
Characters (no spaces)
0
Lines
0
Not saved yet
Converted result
The converted text appears here.Changed words
0
Preserved acronyms
0
Processing mode
LocalIn your browser
Transformation library
Everyday prose conversions for body copy, captions and notes.
Sentence casewriting
Capitalises the first word of each sentence and lowercases the rest. Abbreviations, decimals and initials do not end a sentence.
Scope
Everything, treated as continuous prose.
Protected words / acronyms
These terms will stay unchanged during conversion, on top of 319 built-in acronyms and brand names.
- Runs locally
- No account required
- Draft recovery
- Deterministic, not AI
Simple process
How the Case Converter works
01 Add your text
Type, paste, or drop a TXT, Markdown or CSV file straight onto the editor. Files are read on your device — nothing is uploaded. Your draft is autosaved to this browser as you work.
02 Pick a format
Thirty formats grouped into Writing, Titles, Developer, Creative and Cleanup. Selecting one converts immediately and shows the whole result beside the editor, not in a card.
03 Set the rules
For titles, choose AP, APA, Chicago or MLA and adjust the first-word, last-word, hyphen and subtitle rules. For everything else, choose whether the conversion runs on the document, the selection, each line, or headings.
04 Protect what must not change
A built-in dictionary of 319 acronyms and brand names keeps API, iPhone and NASA intact, plus any term you add. URLs and email addresses are preserved by default.
05 Check what changed
Compare puts the original and the converted text side by side and marks capitalisation-only differences separately from real ones, because those are the changes the eye slides straight over.
06 Export
Copy, download as TXT, Markdown or CSV, replace the source and keep working, or send a converted CSV column back out as a file.
Built for real copy
Case is easy.Names and acronyms are not.
Changing case stops being trivial the moment the text contains a brand, an initialism, a link, or a word a style guide insists stays lowercase.
Start here
Six jobs people actually arrive with.
Changing case is easy until the text contains a name, an acronym, a link or a word a style guide insists stays lowercase. These are the jobs people actually arrive with, and what the studio does about each of them.
Titles
There is no such thing as “title case”.
AP, APA, Chicago and MLA are separate conventions that disagree in public, and the difference shows up in the first line. Every row below was produced by the converter on this page, when this page rendered.
Source
a well-known guide to e-mail and search: what works with ai in 2026
- APAssociated Press Stylebook
A Well-Known Guide to E-Mail and Search: What Works With Ai in 2026
Capitalises words of four letters or more, even prepositions. The newsroom default.
- APAPublication Manual of the APA, 7th edition
A Well-Known Guide to E-Mail and Search: What Works With Ai in 2026
Like AP on length, but capitalises BOTH halves of a hyphenated compound.
- ChicagoChicago Manual of Style, 17th edition
A Well-Known Guide to E-mail and Search: What Works with Ai in 2026
Lowercases every preposition however long, and respects prefixes in hyphenated words.
- MLAMLA Handbook, 9th edition
A Well-Known Guide to E-Mail and Search: What Works with Ai in 2026
Lowercases articles, prepositions and coordinating conjunctions at any length.
| Rule | AP | APA | Chicago | MLA |
|---|---|---|---|---|
| Short prepositions (of, in, to) | lowercase | lowercase | lowercase | lowercase |
| Long prepositions (with, between, through) | Capitalised | Capitalised | lowercase | lowercase |
| Second half of a hyphenated compound | Capitalised unless minor | Always capitalised | lowercase after a prefix | Capitalised unless minor |
| First word after a colon | Capitalised | Capitalised | Capitalised | Capitalised |
| Last word, whatever it is | Capitalised | Capitalised | Capitalised | Capitalised |
Every guide leaves cases to the editor — a proper noun that looks like a preposition, a phrasal verb whose particle should be raised. The studio shows which rule fired for each word, so you can disagree with it deliberately rather than by accident.
Developer
Which identifier case belongs where.
Every row is the same source — “user profile image URL” — run through the converter. Note that the acronym is detected as one unit rather than three letters, which is the difference between a usable name and an unreadable one.
| Convention | Result | Where it belongs |
|---|---|---|
| camelCase | userProfileImageUrl | JavaScript and Java variables, JSON keys in most APIs |
| PascalCase | UserProfileImageUrl | Classes, types, React components, C# methods |
| snake_case | user_profile_image_url | Python, Ruby, Rust, and SQL column names |
| SCREAMING_SNAKE_CASE | USER_PROFILE_IMAGE_URL | Environment variables and compile-time constants |
| kebab-case | user-profile-image-url | CSS classes, URL slugs, npm package names, HTML attributes |
| dot.case | user.profile.image.url | Config keys, namespaces, feature flags, i18n keys |
| path/case | user/profile/image/url | Route segments and file paths |
| Header-Case | User-Profile-Image-Url | HTTP headers: Content-Type, X-Request-Id |
| COBOL-CASE | USER-PROFILE-IMAGE-URL | COBOL identifiers, and some mainframe exports |
| flatcase | userprofileimageurl | Go package names, some database identifiers |
Developer formats treat each LINE as one identifier, so a pasted list of twenty names becomes twenty names rather than one very long one. Digit boundaries are deliberately not split — html5 stays one word, because that is what everyone who writes it means.
Protection
What survives a conversion, and why.
A dictionary of 319 acronyms and brand names is lifted out of the text before anything is converted, then put back in its canonical spelling. URLs and email addresses are handled the same way. These are real before-and-after pairs, not illustrations.
AP Title Case
the api returns json, not xml
The API Returns JSON, Not XML
Registered acronyms keep their capitals.
Sentence case
I LOVE MY IPHONE AND MY MACOS LAPTOP
I love my iPhone and my macOS laptop
Brand spellings are restored, not merely preserved.
lowercase
read https://example.com/API/Guide now
read https://example.com/API/Guide now
A case-sensitive path is not flattened.
UPPERCASE
mail SAM@Example.com about it
MAIL SAM@Example.com ABOUT IT
An address is handed back byte for byte.
Chicago Title Case
this is for all of us
This Is for All of Us
“us” is a pronoun here, not the initialism.
Chicago Title Case
the US economy grew
The US Economy Grew
Typed in capitals, it is the initialism.
The ambiguity rule
A term of two characters or fewer, or one that is also an ordinary English word, only counts when you type it the way the dictionary spells it. US is the initialism; us is the pronoun and is left alone. Without that rule, “this is for all of us” becomes “this is for all of US”, which is the single most common way an acronym dictionary makes output worse rather than better.
Under the hood
How a conversion actually runs.
Four stages, the same four for all 32 formats. Knowing them explains every edge case on this page — and the 17 cleanup actions run in the same place, which is why they share one undo stack.
- 1Protect
URLs, email addresses and dictionary terms are lifted out and replaced with a placeholder no transformation can see inside. That is why a link keeps its capitals while the sentence around it is lowercased.
- 2Tokenize
Prose splits on whitespace and punctuation, keeping apostrophes inside words so doesn’t stays one token. Identifiers additionally split on case runs, so userProfileImageURL becomes four words and not seven.
- 3Apply
Simple formats are a direct mapping. Title case is not: each guide has its own minor-word list, length threshold, hyphen policy and colon rule, and every decision is recorded with the rule that made it.
- 4Restore
The protected spans go back in their canonical spelling, and the result is compared with the source so the changed-word and capitalisation-only counts describe exactly what you are looking at.
Whitespace
- Trim whitespaceRemoves leading and trailing spaces from every line, and blank lines from the start and end of the text.
- Collapse repeated spacesTurns runs of two or more spaces or tabs into one space. Line breaks are untouched.
- Remove duplicate blank linesCollapses three or more consecutive newlines into a single blank line, keeping paragraph breaks.
- Remove every blank lineDrops all empty lines, closing a list up completely.
- Normalise line endingsConverts Windows and classic Mac line endings to plain newlines.
- Convert tabs to spacesReplaces each tab with spaces at the configured width.
- Convert line breaks to spacesUnwraps hard-wrapped text into one paragraph — the fix for copy pasted out of an email.
Characters
- Normalise quotes and dashesCurly quotes become straight, en and em dashes become hyphens, and an ellipsis becomes three dots.
- Strip invisible charactersRemoves zero-width spaces, byte-order marks and control characters, and converts non-breaking spaces.
- Remove HTMLStrips tags and decodes the common entities. A formatting convenience, not a sanitiser.
- Remove punctuationKeeps letters, digits and whitespace. URLs and email addresses are preserved whole.
- Remove diacriticsReduces accented letters to their base form. Changes meaning in some languages.
- Remove emojiRemoves emoji and their whole cluster, including skin tones, flags and joiner sequences.
Lines
- Remove duplicate linesKeeps the first occurrence of each line, comparing without leading or trailing spaces.
- Sort lines A→ZSorts with numeric collation, so item2 comes before item10.
- Sort lines Z→AThe same collation, reversed.
- Reverse line orderPuts the last line first, without changing any line's contents.
Boundaries
Edge cases and limitations.
The places where this tool is a well-informed first pass rather than an answer. Stated here rather than discovered later.
Title case is a judgement, not an algorithm
The four guides follow their published rules, but every guide leaves cases to the editor: a proper noun that looks like a preposition, a phrasal verb whose particle should be raised, a title quoting another title. Treat the output as a well-informed first pass.
Sentence detection is heuristic
Abbreviations outside the built-in list, unusual punctuation and inline citations can put a capital in the wrong place. Titles such as Dr. never end a sentence; general abbreviations such as etc. are decided by what follows them.
Developer formats are for identifiers
They split on case runs, which is right for a name and wrong for a paragraph containing one. Each line is treated as one identifier, so use them on lists rather than prose.
Removing diacritics changes meaning
It is a Unicode normalisation, not a translation. In languages where an accent distinguishes words, the result is a different word.
Very large documents stop previewing live
Past 200k characters the preview converts on demand instead of on every keystroke, and says so. Imported files are capped at 5MB and 50,000 rows.
Remove HTML is not a sanitiser
It strips tags and decodes the common entities as a formatting convenience. Do not use it as a security boundary for untrusted markup.
- No upload
- No account
- No AI API
- Draft stays local
Privacy by design
Nothing you type leaves this page.
There is no submit step and no request that carries your text. That is not a policy statement — it is what the code does, and it is why this tool is usable on material you could not paste into a hosted service.
Nothing is uploaded
Conversion, cleanup, comparison, import and export all run in your browser. There is no submit step and no request that carries your text.
Files are parsed on your device
A TXT, Markdown or CSV file is read in the page. It never reaches a server, and neither does its name.
Your work stays in this browser
Drafts, recent texts, saved pipelines and dictionaries are written to local storage so an accidental tab close does not lose them. They are never synced, and each one can be deleted individually.
One control clears everything
Local history has a single button that deletes every key this tool has written, including the draft.
Safe for text you cannot share
An unpublished headline, an internal roadmap, a client's copy, a column of customer records — none of it leaves the machine you typed it on.
FAQ
Frequently asked questions
Keep working
More text tools
Longer write-ups
Why the style guides disagree
Got a headline that needs a style guide?
Paste it above and compare all four conventions on your own text — with every changed word marked.