Korean Text Pattern Review
This tool flags independent 되요, repeated spaces between Hangul, and spaces before punctuation by default. Adjacent repeated eojeol and text attached after punctuation are optional because the intended style depends on context. Inspect each candidate’s range, surrounding text and proposed edit before applying or rejecting it; decisions can be undone. The tool does not understand sentence meaning or assess overall spelling and grammar.
Key features
- Default: review independent 되요→돼요 with a National Institute of Korean Language reference
- Default: flag repeated spaces between Hangul and spaces before punctuation for contextual review
- Optional: inspect adjacent repeated eojeol and text attached after punctuation
- See each source span, UTF-16 and code point offsets, and surrounding context
- Apply or reject each edit manually, undo up to 50 decisions, and restore the original
- Review up to 10,000 UTF-16 source units in the browser and download revised TXT
How to use
- Load the sample or enter Korean source text.
- Review the three default rules and opt in to the two context-sensitive rules only if needed.
- Find candidates, then inspect each source range, context and reason.
- Apply only the edits that fit the context; reject others. Undo a decision or restore the original if needed.
- Read the revised text again and download a TXT file if useful.
Use cases
- Find an independent 되요 form and check the proposed spelling
- Review repeated spaces and spaces before punctuation in context
- Opt in to repeated-word suggestions and decide whether repetition is intentional
- Keep quoted examples and code while checking source offsets and restoration
Frequently asked questions
Does this check all Korean spelling and grammar?
No. It searches only three default and two opt-in surface patterns. It does not analyze morphology, meaning or honorifics, so it can miss errors and make unsuitable suggestions. You must review the final text.
Has its accuracy or false-positive rate been validated?
The rule behavior and undo flow were tested, but there is no independent benchmark proving precision or false-positive rates across real Korean writing. Nothing is applied automatically; decide from the context of each candidate.
What supports the 되요 suggestion?
The National Institute of Korean Language explains 돼요 as the contracted form of 되어요. This tool only flags the independent eojeol 되요 and does not guess at other 되/돼 contexts.
Should every repeated word be removed?
No. Repetition can express emphasis or sound, so this rule is off by default. Even when you enable it, judge each suggestion from its context.
How are quotations, code and URLs handled?
Candidates inside paired straight or Korean quotation marks, backtick code and HTTP(S) URLs are skipped. Unmatched quotation marks and complex Markdown are not fully parsed.
Are offsets the number of visible characters?
No. UTF-16 offsets are half-open JavaScript string code unit positions. Code point positions are shown too. After emoji or combining characters these counts can differ from visible characters.
Is text sent to a server?
No. Finding candidates, applying, undoing and TXT generation stay in browser memory. The downloaded file contains the entire revised text, so review it before sharing.
Privacy
Original and revised text remain in browser memory and are not sent to the server. The TXT download contains the whole revised text.
Comments & questions