Contraction Expander
Expand contractions for formal academic writing: "don't" becomes "do not", "won't" becomes "will not", "it's" becomes "it is". Or flip the direction and introduce contractions to make stiff AI output read like a human wrote it. Both modes run in your browser.0 replacements
Why contractions matter for writing
Contractions are where voice lives in English. Human writers use "don't", "it's", "they're", "we'd" routinely in casual prose, while academic and formal registers use them far less. Large language models, regardless of register, tend to underuse contractions, so the output reads consistently formal even when the content is casual. That is one of the reasons LLM prose feels stilted.That rate gap is also useful information for AI detectors. Classifiers have learned that contraction frequency correlates with authorship, and training sets include enough paired human-vs-AI samples to make the signal usable. Adding contractions does not fool a detector on its own, but it moves one feature closer to human norms.When to expand
Most academic style guides discourage contractions in formal writing. APA Publication Manual 7th edition allows them but recommends restraint. MLA is similar. Chicago avoids them in formal expository prose. University essay rubrics, especially in philosophy, law, and the humanities, often penalize them outright. If you drafted a paper casually and the final version needs to be formal, expand the contractions before submitting.Expansion is also useful for legal, medical, and technical writing where precision and unambiguity are prioritized over flow. A contract that reads "the party won't be liable" is legally equivalent to "the party will not be liable", but the second form is standard in contract drafting because there is zero ambiguity about whether "won't" is a contraction or a typo for "want."When to contract
When you want text to read natural, conversational, and human. This is the humanization use case. If you pasted an LLM response and it reads stiff, running "contract" mode often lifts the voice significantly: "It is important to note that I am committed" becomes "It's important to note that I'm committed," which immediately sounds less like a consultant and more like a person.Contraction is also the right call for blog posts, email, chat threads, marketing copy, personal essays, social content, newsletters, and roughly any first-person writing where voice matters. If your goal is a friendly, approachable register, the contracted form is almost always closer to what you actually want.One caveat: if the input is already heavily contracted, running "contract" again will not add more. The tool only introduces contractions in places where the expanded form ("do not", "will not", "it is", etc.) actually appears. It does not invent contractions that were not grammatically available in the input.What the tool does and does not cover
The tool handles the complete set of standard English contractions: negations (don't, won't, can't, shouldn't, etc.), pronoun-auxiliary forms (I'm, you're, he's, they'd, we'll), question-word forms (what's, who's, where's), and the common auxiliary-have combinations (would've, could've, should've). It preserves capitalization across all three common cases: lowercase, sentence-case, and ALL CAPS. It also accepts curly apostrophes (U+2019) as equivalent to straight single quotes on input, and outputs only straight ASCII apostrophes.It does not cover regional informal forms like "y'all", "gonna", "wanna", or "gotta" because their expansions are context-dependent and mostly stylistic. It leaves ambiguous possessives alone: "John's book" stays as-is even though "John's" could theoretically be a contraction of "John is." If a possessive-form contraction appears after an actual pronoun ("he's", "she's", "it's"), that form is unambiguous and gets expanded correctly.Everything runs in your browser. The page loads once and then processes each edit locally. No uploads, no logs, no persistence. Close the tab and the text is gone.How this fits a humanization workflow
If you are trying to make AI output read as human, the contraction pass is a useful first step but nowhere near sufficient. Run it, then pass the result through Leap's humanizer for a broader cleanup pass: replacing stock AI phrases, removing em dashes and invisible characters, and flagging sentences of uniform length so you can vary the rhythm. Verify the final result with Leap's free AI score checker to see how the writing signals changed.If you want to pair this with other cleanup passes, see the AI text formatter for smart-quote and formatting cleanup, and the invisible character remover for zero-width Unicode.Rewrite a text in your own wordsPaste it to see what a rewrite costs. Nothing is sent until you choose to.
0 words
Frequently asked questions
Most academic style guides (APA, MLA, Chicago, and many university-level essay rubrics) discourage or prohibit contractions in formal writing. If you drafted with contractions but the final version needs to be formal, expanding them is faster than rewriting by hand, and easier to proofread because every change is one word, not a restructured sentence.
Large language models systematically underuse contractions compared to human writers, so their output reads formal even when the content is casual. That gap is one of the statistical signals AI detectors learn to read. Introducing contractions in places where they are grammatically natural makes the text read human and reduces one of those signals.
No. Contractions are one signal among many: sentence length variation, vocabulary, stock phrases, and punctuation patterns all contribute to a detection score. Adding contractions helps but does not by itself turn flagged text into human-sounding text. For a fuller cleanup, run the output through Leap's humanizer at /humanize. No tool can guarantee passing any detector.
Yes. Possessives like "John's book" are left alone because they don't match contraction patterns. We only fire on unambiguous pronoun-plus-auxiliary-plus-negation forms ("don't", "won't", "it's", "they're"). Ambiguous phrases that could be either a contraction or a possessive are preserved unchanged so you can decide case by case.
Yes. "Don't" becomes "Do not" (sentence case preserved), "DON'T" becomes "DO NOT" (all-caps preserved), and "don't" stays lowercase when expanded to "do not". The same logic runs in reverse when introducing contractions.
We include "ain't" in the contraction set (expands to "is not"). Regional forms like "y'all" and informal constructions like "gonna" or "wanna" are intentionally left out because their expansions are context-dependent and often stylistic rather than grammatical. Handle those by hand if they appear in your text.
Yes. Pick "Expand" to convert contractions to their full forms, or "Contract" to introduce contractions into prose that lacks them. The tool is roughly bidirectional for the common cases: running expand then contract on typical English prose yields something very close to the original.