Grammar Checker

How to Use the Grammar Checker

1. Paste your draft into the input field. The checker works on anything from a single paragraph up to a long-form blog post.
2. Click Check Grammar. The tool scans the text and highlights findings in the output panel, colored by severity: red for errors, yellow for warnings, blue for suggestions.
3. Hover or tap on each highlight to see a short explanation of why the passage was flagged.
4. Fix the text in place by editing the original input; the highlights update when you rerun the check.
5. Copy the cleaned result with the Copy button once you are satisfied, then paste it into your CMS, email client, or document.

What the Checker Catches and How

Everything the checker flags comes from deterministic pattern matching. Double-space detection is a single regex (/ {2,}/g) that also highlights runs of three or four spaces. Repeated words use a case-insensitive pattern like /\\b(\\w+)\\s+\\1\\b/gi, which catches "the the" and "and and" without worrying about case. Capitalization after sentence terminators looks for [.?!]\\s+[a-z] and flags the lowercase starter. Common misspellings are checked against an internal list of the 200 most frequently misspelled English words, including "accomodate" (for "accommodate"), "seperate" (for "separate"), and "recieve" (for "receive"). Passive-voice detection looks for forms of "to be" followed by a past participle, such as was written or are being considered.

The checker intentionally avoids heuristics that require a real parse tree. Subject-verb agreement, article usage ("a" vs. "an"), and dangling modifiers need syntactic analysis, and a rule-based tool that guesses at them produces more noise than signal. For those deeper checks, a trained NLP model like LanguageTool Premium or Grammarly is the right tool. The advantage of the browser-only approach here is that it never sends your draft to a server and it flags the errors that are easiest to miss when re-reading your own writing.

Scenarios Where a Quick Checker Earns Its Keep

• Final pass on a blog post before publishing, when you have read it five times and cannot see the double space you keep retyping between paragraphs.
• Cleaning up a dictation export from Otter.ai or Rev where the speech-to-text engine tends to drop capitalization after periods.
• Reviewing a colleague's release note draft in a hurry before sending it to the customer-facing newsletter.
• Teaching intro writing classes where the instructor wants students to notice passive voice without reading every line by hand.
• Cleaning up a copy-paste from a PDF that introduced random double spaces around line breaks.
• Pre-checking form submissions in a client portal where you want to avoid the obvious errors that make an application look sloppy.

Where Rule-Based Checking Falls Short

The biggest class of false positives is legitimate repeated-word usage. "That that" in "I know that that is true" is grammatical and the checker will flag it. "Had had" in past-perfect constructions ("She had had lunch") is also legitimate. False negatives are also common: the checker cannot tell "affect" from "effect", "then" from "than", or "its" from "it's", because all of those spellings are valid words. Subject-verb disagreement ("The team are winning") and tense inconsistency across paragraphs are completely invisible to a rule-based scanner.

Passive voice flagging also produces noise. Scientific and legal writing intentionally uses passive voice to emphasize the action over the actor ("the sample was heated to 90°C"), and journalism sometimes uses passive voice to omit a subject you do not want to name. Treat passive-voice flags as prompts to reread, not as errors. A human editor should always have the last word.

Grammar Standards and Where Rules Come From

American English grammar conventions in this tool follow the Chicago Manual of Style and the AP Stylebook, which cover business and journalism. British conventions differ in areas like quotation-inside-punctuation and serial commas; the checker does not force Oxford-comma rules. For misspellings, the wordlist is a subset of Wiktionary's "common misspellings" page, cross-referenced with the Oxford English Dictionary usage notes. Sentence-boundary detection follows Unicode Technical Standard #29 in spirit (periods followed by whitespace are boundaries), though the tool does not implement the full UAX #29 rules for abbreviations. Passive-voice identification follows the classic Strunk and White definition, which is more conservative than some modern style guides that tolerate passive voice when it serves clarity.

Comparison With Grammarly, LanguageTool, and Editors

Grammarly is the market leader for a reason: a trained neural model catches subject-verb, tense, and register issues that no rule-based tool can match, and it offers tone suggestions for formal vs. casual writing. The tradeoff is that Grammarly uploads your draft to its servers, requires an account, and costs money for the premium tier. LanguageTool is open-source, can be self-hosted, and supports over 25 languages, but its rule set is heavier and the UI is less polished. Microsoft Editor in Word and Outlook is solid for office documents and it benefits from Microsoft's NLP investment. This browser tool sits at the lightest end: it catches the handful of errors that are common enough to write simple rules for, it never phones home, and it takes two clicks to use. Pair it with one heavier tool for serious work.