Case converter: choose a style and check the whole result
Admin ·
Paste or open plain text, choose its text language and case style, then press Convert case. ToolMellow keeps the input and a separate read-only result so you can inspect the conversion before copying, downloading or using it as the next input.
Choose a style around the text’s destination: prose, a heading or a naming convention. Mechanical case rules do not recognize every proper name, editorial convention or programming-language restriction. Review the whole result, especially when punctuation is removed or a Unicode letter expands.
Convert the whole input, then inspect the separate result
The initial style is Title Case and Keep existing ACRONYMS starts off. Conversion is manual: changing the input or conversion options invalidates the previous result, and you press Convert case again to produce a new one. Empty or whitespace-only input does not enable conversion.
A selection inside the input does not limit conversion. The selected style receives the whole editor, and the output appears in its own read-only editor. Wrap lines changes the display without changing the conversion. Save the result you need before clearing or leaving the page.
- Type, paste or open a UTF-8 text file.
- Choose the text language independently of the interface language.
- Choose a case style and, for prose styles, decide whether to keep existing acronyms.
- Press Convert case to process the whole input.
- Read the result, including names, punctuation and any changed length.
- Copy or download the result, or use it as input for another conversion.
Choose the text language before comparing case styles
The text-language selector starts from the interface language, but you can choose English, Simplified Chinese, Hindi, Spanish, French, Arabic, Bengali, Portuguese, Indonesian, Urdu or German independently. This choice controls locale-sensitive case mappings; it does not translate text or automatically identify every passage’s language.
English text language offers thirteen styles. The other ten choices offer twelve because Headline is English-only. Changing away from English while Headline is selected resets the style to Title. Scripts without letter case may remain visibly unchanged in prose styles, while mixed-script text can still contain letters with upper and lower forms.
Compare the thirteen styles with one English input
With text language set to English, Keep existing ACRONYMS off and input hello WORLD, the styles produce the following results. The same simple input makes their different purposes visible; more complex punctuation, acronyms and Unicode text need the rules described below.
Title, Headline and Sentence work on prose-like words. Upper and Lower map the whole string. The six identifier styles split and join word groups, discarding punctuation and spacing. Alternating and Inverse use individual code points, so they have different behavior from whole-string case mapping.
| Style | Result for hello WORLD | Typical purpose |
|---|---|---|
| Title | Hello World | Capitalize each matched word |
| Headline (English) | Hello World | A heading with a fixed minor-word rule |
| Sentence | Hello world | Capitalize supported sentence starts |
| Upper | HELLO WORLD | All uppercase |
| Lower | hello world | All lowercase |
| Camel | helloWorld | Join groups with later initials uppercase |
| Pascal | HelloWorld | Join groups with every initial uppercase |
| Snake | hello_world | Join groups with underscores |
| Kebab | hello-world | Join groups with hyphens |
| Constant | HELLO_WORLD | Uppercase groups joined by underscores |
| Dot | hello.world | Join groups with dots |
| Alternating | hElLo WoRlD | Alternate letter positions across the text |
| Inverse | HELLO world | Switch each code point’s case |
Separate capitalization of every word from English headline rules
Title finds words beginning with a Unicode letter and continuing through letters, combining marks, numbers or straight/curly apostrophes. It normally lowercases each matched word and uppercases its first code point. It does not preserve the intended spelling of every mixed-case brand or personal name.
English Headline uses the same matched words, but leaves these interior minor words lowercase: a an the and but or nor for so yet as at by in of on per to via. The first and last matched words are capitalized, as is a minor word immediately after a colon. This fixed rule does not establish AP, Chicago or another publisher’s style compliance.
Review sentence starts, names and abbreviations
Sentence normally lowercases matched words, then capitalizes a letter at the start, after a newline, or after ., ! or ? followed by whitespace. A limited set of leading spaces, quotes and opening parentheses is allowed. This punctuation rule is separate from the browser’s linguistic sentence segmentation.
A period without following whitespace does not create the same boundary, and punctuation such as the Hindi danda is not a sentence trigger here. Names, mixed-case brands, abbreviations and an internal English pronoun I need manual review. A mechanical sentence conversion cannot decide every grammatical or editorial boundary.
Keep existing acronyms only where the option applies
Keep existing ACRONYMS is available for Title, English Headline and Sentence. It preserves matched tokens that begin with an uppercase letter and otherwise contain only uppercase letters or Unicode numbers, with at least two characters in total, such as NASA, API2 or API½. The number category includes forms beyond decimal digits. Enable it before converting when those existing forms should remain.
A single uppercase letter and mixed-case HTTPResponse do not satisfy that preservation rule. The option does not discover acronyms or infer their preferred spelling. It does not affect Upper, Lower, the identifier styles, Alternating or Inverse.
Split word groups before joining camel, Pascal, snake, kebab, constant or dot case
The identifier styles split an uppercase acronym before a following capital-plus-lowercase word, and split a lowercase letter or ASCII digit before a following uppercase letter. For example, HTTPResponse becomes the groups HTTP and Response, producing httpResponse, HttpResponse or http_response in Camel, Pascal or Snake.
They then extract Unicode letter/mark/number groups and lowercase them before joining or capitalizing. Punctuation and spacing are discarded as separators; no translation or transliteration is performed. A leading number or retained Unicode letter may be unsuitable for your destination. Check its identifier, URL or domain-name rules separately.
Alternating counts letters globally; Inverse is not a reversible encoding
Alternating starts with a lowercase letter and advances one shared index for every Unicode letter across the whole text. Spaces, punctuation, numbers, emoji and combining marks do not advance it. Letters in scripts without upper/lower forms still advance that index, even when their appearance stays unchanged.
Inverse examines each code point and chooses its lower or upper mapping according to its comparison with uppercase. It can differ from a context-sensitive whole-string transformation. Applying it twice need not restore the original: for example, ß becomes SS, then ss. Keep the original text when you may need its spelling or capitalization later.
ECMA-402: locale-sensitive string case transformationsECMA-262: string case conversion and Unicode mapping
Allow for Unicode expansion and unchanged normalization
Upper and Lower apply locale mappings to the whole string. A mapping can change the number of code points, UTF-16 units and encoded bytes. The converter does not apply Unicode normalization, so it does not deliberately compose a base letter plus combining mark into a precomposed letter.
With English selected, these examples show why case conversion and length are separate questions. The decomposed accent remains decomposed, and an emoji has no case. These fixed examples do not predict every browser’s Unicode data or every language-specific spelling.
| Input | Upper result | What to check |
|---|---|---|
ß | SS | One input letter expands to two output letters |
ffi | FFI | A ligature expands to three ASCII letters |
é (e plus combining accent) | É (E plus combining accent) | The combining sequence is retained |
🌿 | 🌿 | An emoji has no upper/lower form |
ECMA-402: locale-sensitive string case transformationsECMA-262: string case conversion and Unicode mappingMDN: String.length and UTF-16 code units
Read the output counts and editor byte labels as different measures
The completion notice counts the converted output’s words, visible grapheme characters and hard lines. Words use the selected language’s Intl.Segmenter word boundaries, with a supplementary letter/combining-mark rule; graphemes include whitespace. The input and result editors also show their UTF-8 byte sizes. Input limits use UTF-16 units instead.
For the converted output É 🌿, where E is followed by a combining acute accent, the measures below are different. It contains one space and no hard break. A visual line wrap does not add a hard line; the completion notice’s three totals do not include every metric in this table.
| Measure | Value for É 🌿 | Where it matters |
|---|---|---|
| Words | 1 | Converted-output completion notice |
| Visible grapheme characters | 3 | Converted-output completion notice |
| Hard lines | 1 | Converted-output completion notice |
| UTF-16 code units | 5 | Input-length checks |
| UTF-8 bytes | 8 | Editor size labels and exported text bytes |
MDN: Intl.Segmenter and language-sensitive text boundariesMDN: String.length and UTF-16 code units
Check input units and file bytes separately
Open text file reads one file as UTF-8 text and replaces the input when accepted. The picker offers TXT and common text-based extensions including JSON, HTML and CSS. Markup remains literal text. This control does not parse a PDF or DOCX document, and Case Converter does not receive the separate reviewed-PDF-text handoff.
The file must be strictly below 4000000 bytes, and its decoded text must fit the inclusive 1000000-UTF-16-unit input cap. An oversized paste is refused whole after subtracting the replaced selection and normalizing pasted CRLF/CR to LF. Native typing has the editor’s input limit. A rejected import or paste keeps the existing input; line endings may normalize through DOM editing.
| Boundary | Allowed amount | Behavior beyond it |
|---|---|---|
| Current input | At most 1000000 UTF-16 units | Oversized native paste is refused whole |
| Imported file bytes | Strictly below 4000000 bytes | File is rejected before text decoding |
| Decoded imported text | At most 1000000 UTF-16 units | Existing input is preserved |
| Use result as input | Result at most 1000000 UTF-16 units | Original input and generated result remain |
| Find / replacement fields | At most 1000 / 10000 UTF-16 units | Each field has its own editor limit |
MDN: String.length and UTF-16 code unitsMDN: Blob.text() and UTF-8 text decoding
Export the whole converted result before reusing it
Copy result and Download use the whole generated result, regardless of a selection in the input. Download creates UTF-8 plain text named toolmellow-result.txt. It does not save a statistics report or document formatting. If clipboard access fails, follow the page’s manual selection-copy guidance.
The generated output can exceed the input cap because Unicode mappings can expand. Copy and Download keep that complete output. Use result as input checks the input cap separately and refuses a result above it while preserving both editors. For example, 333334 ffi ligatures can produce 1000002 ASCII letters as repeated FFI; export that result or prepare smaller input batches.
Use literal replacement and bounded in-page undo
Find & replace in input searches literal text across the whole input, with optional case matching. A dot is a literal dot, and replacement dollar signs remain literal. Replace all is not limited to a selected passage. Expansion beyond the input cap is refused without replacing the existing input.
Undo and Redo cover accepted input edits, imports, replacements and Use result as input, while conversion options and result generation are separate. Shortcuts are Ctrl/Cmd+Z, Ctrl/Cmd+Shift+Z and Ctrl+Y. Consecutive typing can be grouped. Past and future share at most twenty older snapshots and four million UTF-16 units, excluding current input; older states can be discarded.
Save before Clear, refresh or leaving the page
Clear removes input, generated result, undo/redo history and find/replacement text. It retains conversion settings such as the selected style and text language. Clearing is not an undoable reset, and this editor does not provide durable autosave or recovery after navigation or refresh.
The inspected workflow processes text in a local worker. The page and its assets can still load over the network. Cancel terminates the current worker, and input/option revisions prevent an obsolete result from replacing newer state. The job has a sixty-second deadline including startup; the input cap does not promise a completion time on every device.
If conversion fails, keep or export the text you still need and try a smaller passage or a compatible browser. For a PDF or Word document, extract or copy the intended text first and inspect it before conversion. Local case processing does not extract document layout or identify missing text for you.
Check names, destinations and the original text
For prose, read proper names, acronyms, apostrophes and sentence boundaries after conversion. For headings, compare the fixed English minor-word rule with your publication’s requirements. For identifiers, check discarded punctuation, leading numbers and the naming rules of the destination.
Keep an original copy when capitalization, accents or spacing carry meaning. Use Word Counter for fuller text statistics, Text Cleaner for a separate cleanup task and Text Compare to inspect changes. A case-style result is a starting point for review, not a grammar correction or a guarantee that a name is valid everywhere.
Common questions
How do I convert text to uppercase or lowercase?
Choose the text language, select Upper or Lower, then press Convert case. The whole input is mapped into the separate result. Copy result and Download export that generated result; Use result as input replaces the input only if the result fits its cap.
How are title case and headline case different?
Title normally capitalizes the first code point of every matched word after lowercasing it. English Headline keeps a fixed list of interior minor words lowercase, with first-word, last-word and after-colon exceptions. Headline is available only when text language is English and does not certify a publisher’s style guide.
How does sentence case handle names and abbreviations?
Sentence follows start/newline and dot, exclamation or question-mark-plus-whitespace rules. It does not recognize every name or abbreviation and can lowercase an internal English I. Check names, brands and sentence boundaries manually; punctuation without the supported following whitespace can behave differently.
Can I keep NASA and API uppercase?
Enable Keep existing ACRONYMS for Title, English Headline or Sentence. Existing tokens such as NASA, API2 and API½ match its uppercase-letter/Unicode-number rule. Unicode numbers include forms beyond decimal digits. Mixed-case HTTPResponse and a single uppercase letter do not match. The option does not apply to identifier styles or Upper/Lower.
How do camelCase and snake_case split HTTPResponse?
The identifier tokenizer inserts a boundary between the HTTP acronym and Response. Camel produces httpResponse and Snake produces http_response. Pascal, Kebab, Constant and Dot use the same groups with their own capitalization or separators. Punctuation and spacing are discarded.
How do alternating case and inverse case work?
Alternating starts lowercase and advances across Unicode letters throughout the text; uncased letters still advance it, while nonletters do not. Inverse switches each code point according to its upper/lower mapping. Unicode expansion and contextual mappings mean Inverse applied twice may not restore the input.
Does selecting text convert only that selection?
No. Convert case processes the whole input. A selection in the editor does not narrow the conversion or the tool’s result exports. Prepare an extract as the input if you need a passage converted separately.
Can I open a file and download the converted result?
Open text file accepts a file strictly below 4000000 bytes with decoded text at most 1000000 UTF-16 units. It reads UTF-8 plain text rather than parsing document formats. Download saves the whole generated result as UTF-8 toolmellow-result.txt, even when case expansion makes that result longer than the input cap.
Is my text uploaded or saved?
The inspected conversion and editing workflow runs on your device, with text/history in page memory. Page and asset requests can still use the network. No durable editor autosave is provided. Clear removes the text/history/result, and refresh or navigation can lose the work, so copy or download what you need first.
Why does my text stay unchanged?
It may already match the selected style, or it may use a script without uppercase and lowercase forms. Check the selected text language and press Convert case after changing input or options. A visibly unchanged result does not mean the tool translated or interpreted the text.
Why can I download a result that I cannot use as input?
Case mappings can expand the string. Generated output has no separate million-unit cap in this workflow, so whole-result Copy and Download remain available. Use result as input enforces the input’s 1000000-UTF-16-unit cap and refuses a longer result while keeping both editors.
Why is Convert case unavailable or my old result gone?
Conversion requires non-whitespace input. Editing the input or changing conversion options invalidates an earlier result, so run Convert case again. Wrap lines affects presentation. If an import or paste exceeds a size limit, read the notice and keep the existing input.
Can I convert a PDF or Word document directly?
Open text file does not parse PDF or DOCX content, and Case Converter has no direct reviewed-PDF-text handoff. Extract or copy the intended plain text first, inspect it, then supply it here. Formatting, OCR and document-layout recovery are separate tasks.
Can Undo recover text after Clear?
Clear removes the input and its undo/redo history along with the result and find/replacement text. Conversion settings remain. Copy or download needed text before clearing; the in-page history does not promise recovery after Clear, refresh or navigation.