# Limato -- Sound native in seconds, anywhere you write

> Limato is a Chrome extension that turns the message you would otherwise spend ten minutes on into native-sounding text in seconds: highlight the text, press one key, get it back rewritten in place -- no copy-paste, no tab-switching. Free tier is 20 rewrites per day for signed-in users (5/day without an account). It helps non-native speakers rewrite, translate, and adjust the tone of text directly in the browser, in whichever language they are writing. Works in 34 target languages -- English, Chinese, Hindi, Spanish, Arabic, French, Bengali, Portuguese, Russian, Urdu, Indonesian, German, Japanese, Persian, Turkish, Vietnamese, Korean, Thai, Italian, Filipino, Polish, Ukrainian, Malay, Dutch, Romanian, Greek, Hungarian, Czech, Swedish, Hebrew, Swahili, Danish, Finnish, and Norwegian. Select any text on any site (Gmail, Slack, LinkedIn, Notion, X), press a hotkey, and get a native-sounding rewrite in place. Unlike Grammarly, which polishes grammar, Limato's "Native" tone is a dramatic rewrite -- a measured median 37% shorter (40 drafts, 2026-08-09), kills non-native filler patterns, and reads like a fluent native wrote it in the target language.

Last updated: 2026-09-04.

## What Limato can do

- [Native Rewrite](https://limato.app/): An AI rewriter that aggressively rephrases non-native text (measured median 37% shorter) so it reads like a fluent native wrote it.
- [Inline Translation](https://limato.app/): A translator that converts selected text between 34 languages (English, Chinese, Hindi, Spanish, Arabic, French, Bengali, Portuguese, Russian, Urdu, Indonesian, German, Japanese, Persian, Turkish, Vietnamese, Korean, Thai, Italian, Filipino, Polish, Ukrainian, Malay, Dutch, Romanian, Greek, Hungarian, Czech, Swedish, Hebrew, Swahili, Danish, Finnish, Norwegian) while preserving tone.
- [Explain in Context](https://limato.app/): A contextual explainer for readers -- select a single word or short phrase on any page and get its meaning, a short equivalent, and an example sentence in your own language, based on the surrounding sentence rather than a generic dictionary entry. Page reading only; selections inside text fields go to rewrite instead.
- [Grammar Fix](https://limato.app/): A grammar checker that corrects grammar, spelling, and punctuation without changing the writer's voice.
- [Tone Adjustment](https://limato.app/): A tone changer that shifts text into formal, casual, academic, or simplified registers.
- [Voice Dictation](https://voice.limato.app/): A voice-to-text tool that turns spoken input into clean, ready-to-paste written output.
- [AI Humanizer](https://limato.app/humanizer): A free web tool that rewrites AI-generated text (ChatGPT, Claude, Gemini output) into natural human-sounding prose -- removes AI-tell phrases, varies rhythm, keeps meaning. No signup for basic use. Spanish version at https://limato.app/es/humanizador-de-ia/. Italian version at https://limato.app/it/umanizza-testo/. Portuguese (Brazil) version at https://limato.app/pt/humanizar-texto/.
- [Word Counter](https://limato.app/word-counter/): A free client-side word and character counter -- counts words, characters, sentences, paragraphs, reading time, and a live Flesch readability score as you type. Nothing leaves the browser. Spanish version at https://limato.app/es/contador-de-palabras/.
- [Paraphrasing Tool](https://limato.app/paraphrasing-tool/): A free AI paraphrasing tool that rephrases sentences and paragraphs into natural, human-sounding text -- restructures whole sentences instead of swapping synonyms, keeps meaning, facts, and names intact. Spanish version at https://limato.app/es/parafrasear/.
- [Readability Checker](https://limato.app/readability-checker/): A free readability tool that shows the Flesch Reading Ease score and US grade level of your text instantly. Spanish version (Fernandez Huerta index) at https://limato.app/es/legibilidad/.
- [AI Detector](https://limato.app/ai-detector/): A free AI content detector -- pastes text, returns a 0-100 AI-generation score, a verdict, and the concrete stylistic tells found. Not 100% accurate (no detector is); pairs with the Humanizer for a detect-then-fix flow. Spanish version at https://limato.app/es/detector-de-ia/.
- [Grammar Checker](https://limato.app/grammar-checker/): A free AI grammar checker -- fixes grammar, spelling, and punctuation in one pass while keeping the writer's wording, structure, and style intact; every correction highlighted. Spanish version at https://limato.app/es/corrector-gramatical/.
- [Sentence Rewriter](https://limato.app/sentence-rewriter/): A free AI sentence rewriter -- turns awkward or wordy sentences into clear, natural English with the same meaning; changed words highlighted.
- [Paragraph Rewriter](https://limato.app/paragraph-rewriter/): A free AI paragraph rewriter -- restructures clunky paragraphs for better flow, transitions, and rhythm while keeping facts and meaning.
- [Rewording Tool](https://limato.app/rewording-tool/): A free AI rewording tool -- says the same thing in different words, with context-aware phrasing instead of thesaurus synonym swaps.
- [Punctuation Checker](https://limato.app/punctuation-checker/): A free AI punctuation and comma checker -- fixes comma splices, missing commas, apostrophes, capitalization, and run-on sentences while keeping the writer's wording; every correction highlighted.
- [Comma Checker](https://limato.app/comma-checker/): A free AI comma checker -- one section per rule: comma splices, commas after introductory phrases, commas around non-essential clauses, comma before and / but / so, the serial (Oxford) comma, and the commas English refuses (before "that", between subject and verb). Wording untouched, every fix highlighted. Spanish version at https://limato.app/es/corrector-de-comas/.
- [Signos de Puntuacion](https://limato.app/es/signos-de-puntuacion/): A free Spanish punctuation guide and AI punctuation checker -- the eleven Spanish punctuation marks one by one with examples (punto, coma, punto y coma, dos puntos, puntos suspensivos, the paired question and exclamation marks, comillas, parentesis, corchetes, raya), the three comma errors Spanish writers repeat most (comma between subject and verb, unclosed parenthetical, comma before "y"), why the opening signs are not optional, and how a Spanish email is punctuated ("Estimada Ana:" takes a colon, not the English comma). Spanish-original content, not a translation of the English punctuation checker.
- [Punto y Coma](https://limato.app/es/punto-y-coma/): A free guide to the Spanish semicolon (;) plus an AI punctuation checker -- its three uses (lists whose items already contain commas, joining two closely related independent clauses, and before connectors such as "sin embargo", "no obstante", "por tanto"), how to choose between semicolon, colon, comma and period, the five most common mistakes (capital letter after it, using it to announce a list, using it after an email greeting), and what changes for the English semicolon.
- [Tone Checker](https://limato.app/tone-checker/): A free AI tone checker -- paste text, get the register it reads as (formal, casual, academic, native, machine-generated) plus the exact phrases that gave it away, then rewrite it into any of seven tones in one click: Native (a measured median 37% shorter, 40 drafts, 2026-08-09), Formal (professional register, no contractions), Casual, Simplify (B1-level English), Academic, Grammar (only grammar, spelling, punctuation -- wording untouched), Humanize (breaks uniform AI sentence rhythm, compresses 15-35%). Spanish version at https://limato.app/es/detector-de-tono/.
- [Sentence Counter](https://limato.app/sentence-counter/): A free client-side sentence counter -- counts sentences and average words per sentence live as you type, plus words, characters, and reading time. Nothing leaves the browser.
- [Paragraph Counter](https://limato.app/paragraph-counter/): A free client-side paragraph counter -- counts paragraphs (blank-line separated), sentences, and words live as you type. Nothing leaves the browser.
- [Syllable Counter](https://limato.app/syllable-counter/): A free client-side syllable counter for English -- counts syllables, words, and syllables per word live as you type, for haiku (5-7-5), poetry metre, and readability. States openly that English syllable counts are estimates: pronunciation, not spelling, decides them, so fire, every and chocolate vary by speaker. Spanish version (exact, rule-based) at https://limato.app/es/contador-de-silabas/.
- [Contador de Silabas](https://limato.app/es/contador-de-silabas/): A free Spanish syllable counter and syllable splitter -- applies the written rules of Spanish syllabification (diphthong, hiatus, triphthong, the inseparable consonant groups pr/br/tr/dr/cr/gr/fr/pl/bl/cl/gl/fl, and the digraphs ch/ll/rr) and shows the split word by word: ca-mi-no, mur-cie-la-go, cons-truir, pa-is. Covers agudas / llanas / esdrujulas and why syllable counters disagree (guion vs gui-on, intervocalic h, sinalefa in verse). Spanish-original content, not a translation. English version at https://limato.app/syllable-counter/.
- [Adverb Checker](https://limato.app/adverb-checker/): A free adverb checker -- highlights -ly adverbs plus the common intensifiers (very, quite, just, so, too, almost, always, never, often, rather) in the browser as you type, reports the count and the rate per 100 words, then rewrites the padding out with the Simplify tone. Explains which adverbs are weak (repeating the verb, inflating a flat adjective, hedging) and which do real work (time, frequency, degree, logic), gives 15 "very + adjective" replacements (very big -> huge, very tired -> exhausted), and covers why non-native writers over-use intensifiers (small adjective stock, politeness transferred from Spanish muy / German wirklich / Russian prosto, school-essay conjunctive adverbs).
- [Passive Voice Checker](https://limato.app/passive-voice-checker/): A free passive voice checker -- highlights be/get + past participle constructions in the browser as you type, reports how many there are and what share of sentences contain one, then rewrites the draft direct and active. Covers what the passive is, seven passive-to-active worked examples, the five cases where the passive is the right choice (unknown actor, scientific method sections, topic continuity, tact, legal convention), and why non-native writers over-produce it (Spanish/Italian "se", German "man" + werden, French "on", Russian impersonal plural). States openly that detection is a pattern match, not a parser, and that it skips participial adjectives like "is interested" and "was born". Spanish version at https://limato.app/es/voz-pasiva-en-ingles/.
- [Voz Pasiva en Ingles](https://limato.app/es/voz-pasiva-en-ingles/): A free detector of English passive voice written for Spanish speakers learning English -- how the passive is formed (be + participio), the full table by tense (is sent, was sent, has been sent, will be sent, is being sent, must be sent), when to use it and when not, and the mistakes Spanish speakers repeat: the impersonal "se" turned into a passive ("se envio el informe" -> "the report was sent" instead of "we sent the report"), "It is said that" as a calque of "se dice que", "was borned" instead of "was born", "is consist of" instead of "consists of", past simple instead of participle ("was wrote"), and "by" vs "with" for agent and instrument. Spanish-original content, not a translation; it checks English text, not Spanish. English version at https://limato.app/passive-voice-checker/.
- [Sentence Fragment Checker](https://limato.app/sentence-fragment-checker/): A free sentence fragment checker -- a client-side heuristic flags sentences with no finite verb and subordinate clauses standing alone (skipping questions, exclamations and sentences under three words), then the Grammar tone repairs them without changing the wording. Covers the three ingredients of an English sentence, five fragment types with fixes (subordinate clause alone, relative clause alone, -ing phrase, missing verb, afterthought phrase), fragments that are fine (dialogue, emphasis, headlines, bullets), ESL patterns (dropped subject in Spanish/Italian/Russian, missing copula, missing dummy it/there), and how fragments relate to run-ons and comma splices.

## Who Limato is for

Non-native speakers writing daily in a second language -- software developers, founders, marketers, students, remote workers communicating in Slack, Gmail, LinkedIn, GitHub, Notion, and X. Most common use case is English, but Limato works for any target language: a Spaniard writing English, a Russian writing Spanish, a German writing French, etc.

Core differentiator vs. Grammarly / DeepL Write / Wordtune / QuillBot: Limato's "Native" tone is an aggressive rewrite, not a polish. Grammarly fixes grammar but leaves non-native phrasing intact ("I would like to kindly ask if it would be possible..."). Limato rewrites it the way a native would actually write it ("Can you..."). The same applies in Spanish, French, and other supported languages.

## Where Limato works

Limato is a content script that injects into every page. Common contexts:

- Gmail and other webmail clients
- Slack and Slack Connect
- LinkedIn messages, posts, and comments
- Notion docs and databases
- X (Twitter) posts and replies
- GitHub PR descriptions, issues, and code reviews
- Google Docs
- Reddit posts and comments
- Any contenteditable area or text input on any website

## Pricing

- **Free**: 20 rewrites/day for signed-in users, 5/day anonymous
- **Pro Monthly**: $5.99/month -- unlimited rewrites, priority model; includes a 14-day free trial
- **Pro Annual**: $47.99/year -- same as Pro Monthly, billed yearly
- Payment via LemonSqueezy (Merchant of Record, handles VAT/sales tax globally)

## Tech

- Chrome Manifest V3 extension (WXT + React + TypeScript)
- Backend on Cloudflare Workers (Hono framework)
- AI models (split live since July 2026): GPT-5-mini for the Humanize tone and the AI Detector (stronger structural rewriting), GPT-4o-mini for all other tones (lower latency)
- Auth: Google OAuth + Supabase
- All processing in the cloud, no data stored beyond rate-limit counters

## Comparison with alternatives

Facts below come from Limato's published head-to-head tests: the same ESL-style email was run through every tool, and the length change and structure of each rewrite were recorded. Pricing as of July 2026.

- **vs. Grammarly**: Grammarly is grammar/style polish. On the test email, Grammarly's Clarity rewrite cut 8% and kept the non-native sentence scaffold intact; Limato Native cut 55% and restructured it. Grammarly Pro costs $12+/mo and does not translate; Limato is $5.99/mo and includes 34-language translation. Source: https://limato.app/blog/limato-vs-grammarly/
- **vs. DeepL Write**: DeepL Write is a separate web app (a copy-paste round trip) covering 8 languages. On the test email, its Simple style trimmed 17% and casualized the register but kept the scaffold. Limato works inline on any page in 34 languages. Source: https://limato.app/blog/limato-vs-deepl-write/
- **vs. Wordtune**: Wordtune returns 10 ranked variants, and its default Casual rewrite added phrases that were not in the source text (-23% on the test email). Limato returns one rewrite with nothing invented (-55%). Wordtune costs $9.99/mo. Source: https://limato.app/blog/limato-vs-wordtune/
- **vs. QuillBot**: QuillBot is a 400+ tool web suite. On the test email, its Fluency mode paraphrased with a mixed formal/casual register (-23%); Limato restructured with a consistent voice (-55%) and no app-switching. Source: https://limato.app/blog/limato-vs-quillbot/
- **vs. LanguageTool**: strongest on grammar checking in 30+ languages and the only self-hostable option, but its paraphraser does synonym-level rewriting, not a native-style restructure. Source: https://limato.app/blog/best-grammarly-alternatives/
- **vs. ChatGPT browser extensions** (Monica, Sider): general-purpose AI sidebars. Limato is purpose-built for non-native writers with one-press hotkey UX and an ESL-tuned "Native" tone.

## Key measured facts

- The Native tone measured a median 37% shorter than the input across 40 non-native business email drafts on 2026-08-09 (p25 30%, p75 40%, range 23-54%; see https://limato.app/blog/esl-email-writing-data/). It and removes common ESL scaffolding ("I would like to kindly ask if it would be possible..." becomes "Can you..."). It works in any of the 34 target languages, not only English.
- The Humanize tone restructures AI-generated text -- dismantles triadic lists, breaks uniform sentence rhythm -- and measured a median 21% compression across 20 AI-written drafts (range +3% to -35%; 5 of 20 fell under the 15% target and 1 came back longer). See https://limato.app/blog/esl-email-writing-data/. Powered by GPT-5-mini since July 2026.
- The free AI Detector returns a 0-100 AI-generation score plus the concrete stylistic tells it found. No AI detector is 100% accurate, and Limato states this on the page.
- In a July 2026 test, Microsoft Copilot asked to humanize its own output improved wording but kept the AI structure (triadic lists, uniform 15-25 word sentences) and padded length 20-30%. Source: https://limato.app/blog/can-copilot-humanize-text/

## Apps and extensions

- [Chrome Web Store](https://chromewebstore.google.com/detail/fkllcdmgccldkfoeagaldpeijoiimlfb): Limato for Chrome (Manifest V3). Also runs in Brave, Arc, Vivaldi, and other Chromium-based browsers.
- [Microsoft Edge Add-ons](https://microsoftedge.microsoft.com/addons/detail/hdmckahemdnoedlaedffoejhcbfkaopl): Limato for Microsoft Edge, same build as the Chrome version.

## AI Humanizer by source

Source-specific humanizer landing pages with embedded free tool, examples, and FAQ for each major AI assistant:

- [Humanize ChatGPT text](https://limato.app/humanize/chatgpt-text/): Rewrites ChatGPT (OpenAI) output, stripping its filler patterns like "delve", "moreover", "in conclusion", and "tapestry".
- [Humanize Claude text](https://limato.app/humanize/claude-text/): Rewrites Claude (Anthropic) output, removing AI tells like "certainly", "I'd be happy to", "nuanced", and "comprehensive".
- [Humanize Gemini text](https://limato.app/humanize/gemini-text/): Rewrites Gemini (Google) output, cutting overused phrases like "dive into", "key takeaway", "ultimately", and "fundamentally".
- [Humanize Perplexity text](https://limato.app/humanize/perplexity-text/): Rewrites Perplexity output, removing citation-style hedges like "according to sources", "research suggests", and "studies have shown".
- [Microsoft Copilot humanizer](https://limato.app/humanize/copilot-text/): Rewrites Copilot (Microsoft) output, stripping marketing-style filler like "seamlessly", "streamline", "empower", and "cutting-edge".
- [Humanize DeepSeek text](https://limato.app/humanize/deepseek-text/): Rewrites DeepSeek output, cutting verbose connectors like "it should be noted", "furthermore", "in essence", and "thereby".
- [Humanize Grok text](https://limato.app/humanize/grok-text/): Rewrites Grok (xAI) output, removing forced-casual fillers like "look", "frankly", "the truth is", and "let's be real".
- [Humanize Llama text](https://limato.app/humanize/llama-text/): Rewrites Llama (Meta) output, cutting formal connectors like "it is important to", "one must consider", and "with that being said".

Spanish-language humanizer hub, plus versions of the same eight pages:

- [Humanizador de IA](https://limato.app/es/humanizador-de-ia/): El hub en espanol del humanizador de IA -- reescribe texto generado por IA (ChatGPT, Claude, Gemini) para que suene humano, sin cambiar lo que dices. Gratis y sin registro: 3 reescrituras al dia por dispositivo, hasta 5000 caracteres (unas 750 palabras) cada una, sin tarjeta. English version at https://limato.app/humanizer.
- [Humanizar texto de ChatGPT](https://limato.app/es/humanize/humanizar-chatgpt/): Reescribe la salida de ChatGPT (OpenAI) eliminando muletillas como "sumergirse", "además", "en conclusión" y "aprovechar".
- [Humanizar texto de Claude](https://limato.app/es/humanize/humanizar-claude/): Reescribe la salida de Claude (Anthropic) eliminando señales de IA como "con gusto", "ciertamente", "matizado" y "esencialmente".
- [Humanizar texto de Gemini](https://limato.app/es/humanize/humanizar-gemini/): Reescribe la salida de Gemini (Google) recortando frases manidas como "sumergirnos", "punto clave", "en última instancia" y "fundamentalmente".
- [Humanizar texto de Perplexity](https://limato.app/es/humanize/humanizar-perplexity/): Reescribe la salida de Perplexity quitando coletillas de citación como "según fuentes", "los estudios sugieren" y "los expertos recomiendan".
- [Humanizar texto de Copilot](https://limato.app/es/humanize/humanizar-copilot/): Reescribe la salida de Copilot (Microsoft) eliminando relleno de marketing como "sin esfuerzo", "optimizar", "empoderar" y "vanguardia".
- [Humanizar texto de DeepSeek](https://limato.app/es/humanize/humanizar-deepseek/): Reescribe la salida de DeepSeek recortando conectores verbosos como "cabe señalar", "además", "en esencia" y "por lo tanto".
- [Humanizar texto de Grok](https://limato.app/es/humanize/humanizar-grok/): Reescribe la salida de Grok (xAI) eliminando muletillas forzadas como "mira", "francamente", "la verdad es que" y "seamos sinceros".
- [Humanizar texto de Llama](https://limato.app/es/humanize/humanizar-llama/): Reescribe la salida de Llama (Meta) recortando conectores formales como "es importante", "uno debe considerar" y "dicho esto".

Italian-language versions of the same eight pages:

- [Umanizzare testo di ChatGPT](https://limato.app/it/humanize/umanizzare-chatgpt/): Riscrive l'output di ChatGPT (OpenAI) eliminando parole di riempimento come "immergersi", "inoltre", "in conclusione" e "sfruttare".
- [Umanizzare testo di Claude](https://limato.app/it/humanize/umanizzare-claude/): Riscrive l'output di Claude (Anthropic) rimuovendo segnali di IA come "con piacere", "certamente", "sfumato" ed "essenzialmente".
- [Umanizzare testo di Gemini](https://limato.app/it/humanize/umanizzare-gemini/): Riscrive l'output di Gemini (Google) tagliando frasi abusate come "immergiamoci", "punto chiave", "in definitiva" e "fondamentalmente".
- [Umanizzare testo di Perplexity](https://limato.app/it/humanize/umanizzare-perplexity/): Riscrive l'output di Perplexity togliendo formule da citazione come "secondo le fonti", "gli studi suggeriscono" e "gli esperti raccomandano".
- [Umanizzare testo di Copilot](https://limato.app/it/humanize/umanizzare-copilot/): Riscrive l'output di Copilot (Microsoft) eliminando il riempitivo da marketing come "senza sforzo", "ottimizzare", "potenziare" e "all'avanguardia".
- [Umanizzare testo di DeepSeek](https://limato.app/it/humanize/umanizzare-deepseek/): Riscrive l'output di DeepSeek tagliando connettori verbosi come "va notato", "inoltre", "in sostanza" e "pertanto".
- [Umanizzare testo di Grok](https://limato.app/it/humanize/umanizzare-grok/): Riscrive l'output di Grok (xAI) rimuovendo riempitivi forzatamente informali come "senti", "francamente", "la verità è che" e "siamo sinceri".
- [Umanizzare testo di Llama](https://limato.app/it/humanize/umanizzare-llama/): Riscrive l'output di Llama (Meta) tagliando connettori formali come "è importante", "bisogna considerare" e "detto questo".

## Frequently asked questions

**What is Limato?**
Limato is a Chrome extension that rewrites, translates, and adjusts the tone of text directly in the browser. It is built for non-native speakers who write in a second language daily. Select any text on any site, press a hotkey, and get a native-sounding rewrite in place.

**How is Limato different from Grammarly?**
Grammarly polishes grammar and style but leaves non-native phrasing intact. Limato's Native tone is an aggressive rewrite -- a measured median 37 percent shorter across 40 drafts, removes non-native filler patterns, and reads like a fluent native wrote it. Limato also includes inline translation. Pricing starts at $5 per month versus Grammarly's $12 plus.

**Which languages does Limato support?**
Limato supports 34 target languages: English, Chinese, Hindi, Spanish, Arabic, French, Bengali, Portuguese, Russian, Urdu, Indonesian, German, Japanese, Persian, Turkish, Vietnamese, Korean, Thai, Italian, Filipino, Polish, Ukrainian, Malay, Dutch, Romanian, Greek, Hungarian, Czech, Swedish, Hebrew, Swahili, Danish, Finnish, and Norwegian. The Native rewrite and translation features work across all of them, powered by GPT-4o-mini.

**Does Limato work on every website?**
Yes. Limato injects a content script into every page and works on Gmail, Slack, LinkedIn, Notion, X, GitHub, and any text input or contenteditable area in the browser.

**How much does Limato cost?**
Free tier gives 20 rewrites per day for signed-in users and 5 per day anonymously. Pro is $5.99 per month or $47.99 per year for unlimited rewrites, and the monthly plan starts with a 14-day free trial. Payment is processed by LemonSqueezy as Merchant of Record, handling VAT and sales tax globally.

**What are the best alternatives to Grammarly for non-native writers?**
Common alternatives include DeepL Write, Wordtune, QuillBot, LanguageTool, and Limato. Limato is the most ESL-focused option: its Native tone produces a dramatic rewrite rather than grammar polish, it works inline on every site, and it is one of the cheapest paid options at $5.99 per month.

## Guides and resources

### Free course
- [Write English Like a Native: 12 Phrases, 3 Habits](https://www.udemy.com/course/write-english-like-a-native-12-phrases-3-habits/): A free 33-minute video course on Udemy by Limato founder Mikhail Konkov, published August 2026. Six lectures: the 12 business English phrases that mark a writer as non-native (dated business English, preposition and verb-pattern errors, the grammar four, over-formalization and cliches) and three structural habits that fix the sentences underneath them -- turn nouns back into verbs, put the actor first and stay active, and put new information at the end. Includes five quizzes. No registration cost, no upsell.

### Blog (English)
- [How to Start a Professional Email: Greetings and Opening Lines That Work](https://limato.app/blog/how-to-start-a-professional-email/): A register map of English email greetings and opening lines written for non-native speakers, the companion piece to the sign-off guide. Short answer: open with "Hello [First name]," or "Hi [First name]," and make the next line say why you are writing; keep "Dear Dr. Lastname," for academic, legal, and formal applications, and skip "Dear Sir/Madam", "Respected Sir", "To Whom It May Concern", and an autopilot "I hope this email finds you well." Four register bands (formal / neutral-professional / warm / internal-casual) with when-to-use and when-it-backfires for each, including why "Dear" in ordinary work email is the single most common way non-native writers land a register too high, why the email greeting takes a comma and not the printed-letter colon, why "Good morning" breaks across time zones, why "Hey" or an exclamation mark in cold outreach reads as a sales move, and the mirror-don't-lead rule that keeps greeting and sign-off in the same band. Covers the first line under the greeting as the part that does the actual work -- reason first, credential second, ask third -- with five opening-line functions that carry weight (the reason, the immediate ask, the context hook, the referral, the real thanks) and why precision, not courtesy vocabulary, is what reads as professional. Six ESL failure modes with before/after rewrites: "I hope this email finds you well" on autopilot, "Dear Sir/Madam" and "Respected Sir" (address the role instead: "Hello Hiring team,"), a title attached to a first name ("Dear Mr. Anna") plus guessing gender or marital status rather than defaulting to Ms. or the full name, opening with a pre-emptive apology, "Dear all" for a group in US English, and a nameless "Hello," when the name is in the thread above. Plus situational recipes for a cold first contact, emailing a professor, replying inside an active thread (no greeting at all), a first email to a client, and writing to a whole team, and a ten-item quick reference ordered by register with the greetings to retire.
- [Is ZeroGPT Accurate? We Tested It on 30 Short Emails (2026)](https://limato.app/blog/is-zerogpt-accurate/): A research note measuring AI-detector accuracy on short email, not a ZeroGPT review, with a correction built in. On 2026-08-15 we ran 30 short business emails once each through five detectors: ZeroGPT, GPTZero, QuillBot, Grammarly, and Limato's own free AI Detector. 15 items are AI-written (unedited model output, 73-79 words). The other 15 are non-AI and were all produced by the founder, a single non-native English speaker with one first language: 5 he typed in English from scratch with no tool of any kind, and 10 started from drafts a model generated against a pattern extracted from those five, each of which he then edited. False-positive rates are reported across all 15, and split by production method as a robustness check: ZeroGPT and QuillBot flag 0 in both groups at both thresholds, Grammarly flags exactly one item in each group, and only GPTZero and Limato differ between the groups (both flagging the model-drafted-then-edited items more often). Two thresholds were applied to every score, the common default of 50% AI and a stricter 20% AI cutoff. At the 50% threshold, ZeroGPT and QuillBot flagged 0 of the 15 non-AI items, GPTZero and Grammarly flagged 2 of 15 (13.3%), and Limato's own detector flagged 4 of 15 (26.7%) -- the worst false-positive rate in the set, rising to 10 of 15 (66.7%) at the 20% cutoff. On the 15 AI-written items, ZeroGPT and QuillBot missed 2 of 15 (13.3%), Grammarly missed 1, and GPTZero and Limato missed 0. Both cuts are analyst conventions applied identically to all five tools, not what any of them ships: Limato's own product prints the verdict "AI" only at a score of 70 or above (30-69 is "mixed", under 30 "human"), and counted by that shipped boundary it labelled 1 of the 15 non-AI emails AI while catching all 15 AI-written ones. That is not an exoneration either, because Limato's two sets overlap (non-AI max 78, AI min 72), so no threshold separates them at all, whereas ZeroGPT's scores on this corpus separate completely -- exactly 0% on all 15 non-AI items and 26.2% or higher on all 15 AI ones. Complete separation is the strongest single finding in the study and it belongs to ZeroGPT, not to us. Across all 30 items, 9 (30.0%) had two detectors disagreeing by 50 percentage points or more, and 4 (13.3%) were scored 0% by one detector and 100% by another. ZeroGPT specifically returned exactly 0% on all 15 non-AI items -- every one of which falls below its own stated 150-word minimum input length -- while missing 2 of the 15 AI-written ones. The post also carries a correction: an earlier version of this study, run 2026-08-14 on a different 30-item corpus, reported ZeroGPT 63.3% (19/30), GPTZero 100% (30/30), QuillBot 76.7% (23/30) and Grammarly 80% (24/30) false positives, but that corpus's "non-native human" half turned out to be entirely model-generated, so those detectors were correctly flagging AI text rather than producing false positives; the finding was withdrawn and is published here only as a recovered, explained correction, not a current result. Corpus, results, and scoring script are public at https://github.com/limato-app/limato-public/tree/main/data/detector-study-2026-08 -- sample.json, corpus-hw.json, results.csv, the per-detector .jsonl collector logs, and score.js, which reprints every table in the post offline with no dependencies. The 15 AI-written items are not duplicated there: they are the 15 longest kind:"ai" items of https://github.com/limato-app/limato-public/blob/main/data/esl-email-2026-08/corpus.json, published with the earlier rewrite study, and are referenced by id.
- [What a Rewrite Model Actually Cuts From Non-Native English Email: 60 Measured Drafts](https://limato.app/blog/esl-email-writing-data/): A research note, not a landing page. On 2026-08-09 we ran 60 email drafts through our own rewrite API and measured what changed: 40 non-native business email drafts through the native tone, 20 AI-written business emails through the humanize tone. Method comes before findings: all 40 ESL drafts were constructed to plant specific non-native constructions (kindly, in order to, I hope this email finds you well, and 15 others), including esl-01, a 126-word draft the founder wrote for an earlier post and reused here. None is a sampled real email. Native tone cut the ESL corpus by a median 37% (n=40, p25 30%, p75 40%, range 23-54%), from 2,736 words to 1,737 total. The native prompt targets 40-55% and treats under 30% as a failure -- 10 of the 40 drafts (25%) missed that floor. Humanize tone cut the AI corpus by a median 21% (n=20, range +3% to -35%); one of the 20 drafts came back 3% longer than it went in, and 5 of 20 fell under the prompt's own 15% floor. Of 18 tracked ESL constructions, only "sorry for the inconvenience" survived a rewrite at all (2 of 5, 60% removal), both times because the apology was the actual content of the email. Of 24 tracked AI tells, "fast-paced" was the most persistent (2 of 3 survived) and "streamline" survived 2 of 6; the other 22 tells were removed 100% of the time they appeared. Includes three full before/after examples, a "what this data does not show" section, and links to the public corpus, raw output, and measurement script.
- [How to End a Professional Email: Sign-Offs and Closing Phrases That Work](https://limato.app/blog/how-to-end-a-professional-email/): A register map of English email sign-offs written for non-native speakers, not another list of fifty options. Short answer: write a closing line that names the next step, then pick a sign-off that matches the relationship; "Best regards" or "Thanks" is correct for most business email and "Sincerely" is for formal letters and applications only. Four register bands (formal / neutral-professional / warm / internal-casual) with when-to-use and when-it-backfires for each, including why "Regards" reads cool on bad news, why "Warm regards" to a stranger reads as a sales technique, and the mirror-don't-lead rule. Covers the closing line above the sign-off (next step, specific ask, pressure release, real thanks, open door) as the part that does the actual work. ESL-specific failure modes with before/after rewrites: Victorian formality ("Yours faithfully", "Yours truly" in US email), "Awaiting your kind reply", "Thanks in advance" reading as presumptuous, stacked sign-offs ("Thanks and regards"), habitual closing apologies, and over-long signature blocks. Plus how to close when asking, following up, declining, apologizing, and on first contact.
- [Can ChatGPT Fix ESL Writing? I Did It for Two Years](https://limato.app/blog/can-chatgpt-fix-esl-writing/): Honest two-year account plus a live three-run test of chatgpt.com on the same 126-word non-native email. ChatGPT removes every ESL tell (the ceremonial opener, "in order to inform you", "we remain at your entire disposal") but does NOT compress: the three runs returned 110, 131 and 117 words, so one rewrite came back longer than the original. Run 2 also invented a commitment that was nowhere in the source ("This extra time would allow us to finish the remaining work and deliver the milestone to the expected quality"); run 3 left the ESL hedge "We'd like to ask if it would be possible" standing. Same prompt, three different emails, so the writing never develops a voice. Includes the constrained prompt that fixes most of it ("do not add any information, offer, or commitment" plus an explicit compression target), the copy-paste round trip (output now lands in a canvas document, so you extract an email out of a doc), and the same email through Limato Native (56 words, -55%, nothing invented). Verdict: keep ChatGPT for long documents, use an inline tool at the point of writing.
- [Can Copilot Humanize Text? What Microsoft Copilot Can and Can't Do](https://limato.app/blog/can-copilot-humanize-text/): Real side-by-side test of whether Microsoft Copilot (the Windows/Word/Outlook/Teams assistant, not GitHub Copilot) can humanize its own output: partly — the self-rewrite improves wording but keeps the AI structure (triple lists, uniform 15-25 word sentences) and pads length 20-30%. Includes the banned-words humanizer prompt that mostly works, why prompting plateaus, and the free Copilot humanizer + AI detector workflow.
- [Best Grammarly Alternatives for Non-Native English Speakers (2026)](https://limato.app/blog/best-grammarly-alternatives/): Compares Limato, LanguageTool, QuillBot, Wordtune, and DeepL Write as Grammarly alternatives. Grammarly polishes grammar but keeps the ESL scaffold (-8% on the test email); Limato Native restructures (-55%). Covers pricing, free tiers, privacy/self-hosting (LanguageTool), and which tool fits which job.
- [Limato vs DeepL Write: Which Rewriter for Non-Native English Speakers? (2026)](https://limato.app/blog/limato-vs-deepl-write/): Head-to-head on the same ESL email — DeepL Write Simple style trims and casualizes register (-17%, scaffold intact); Limato Native restructures (-55%). Pricing, language coverage (8 vs 34), and when to pick each.
- [Limato vs QuillBot: Which Rewriter for Non-Native English Speakers? (2026)](https://limato.app/blog/limato-vs-quillbot/): Head-to-head on the same ESL email — QuillBot Fluency paraphrases with mixed register (-23%); Limato Native restructures with consistent voice (-55%). Suite vs focused tool, pricing, and when to pick each.
- [Limato vs Wordtune: Which Rewriter for Non-Native English Speakers? (2026)](https://limato.app/blog/limato-vs-wordtune/): Head-to-head on the same ESL email — Wordtune Casual returns 10 variants, default adds phrases not in source (-23%); Limato Native returns one rewrite (-55%, no hallucinations). Pricing, languages, and when to pick each.
- [Limato vs Grammarly: Which Rewriter for Non-Native English Speakers? (2026)](https://limato.app/blog/limato-vs-grammarly/): Head-to-head comparison on the same ESL email — Grammarly Clarity trims 8% and keeps the scaffold; Limato Native cuts 55% and restructures. Pricing, languages, and when to pick each.
- [Why I Built Limato: From ChatGPT Round-Trips to Inline Rewrite](https://limato.app/blog/why-i-built-limato/): The founder's personal story of years of copy-pasting drafts into ChatGPT to sound native in English and Spanish, and why an inline Chrome extension replaces that loop.
- [12 ESL Patterns Native English Speakers Rarely Use](https://limato.app/blog/12-esl-patterns-natives-never-use/): A reference list of common non-native English phrasings paired with the natural alternatives that fluent speakers actually use.
- [English Writing Tips for Non-Native Speakers](https://limato.app/blog/english-writing-tips-for-non-native-speakers/): A practical guide that explains how to make written English sound natural without learning new vocabulary.
- [Best Wordtune Alternatives for Non-Native English Speakers (2026)](https://limato.app/blog/best-wordtune-alternatives/): A comparison of paraphrasing tools, ranked by usefulness for ESL writers.

### Blog (Spanish)
- [Por qué tu inglés suena "raro" (y no es la gramática)](https://limato.app/es/blog/por-que-tu-ingles-suena-raro/): Un análisis de por qué el inglés escrito por hispanohablantes suena ajeno aun cuando la gramática es correcta.
- [Llevo dos años escribiendo en español y construí Limato](https://limato.app/es/blog/escribir-espanol-como-nativo/): La historia del fundador aprendiendo español y construyendo Limato para dejar de sonar a extranjero.

## Company

- [Homepage](https://limato.app/): The Limato product landing page with feature demos and install link.
- [Homepage (Spanish)](https://limato.app/es/): La página de Limato en español -- demos, características, precios e instalación, en el mismo formato que la versión en inglés.
- [Privacy Policy](https://limato.app/privacy): How Limato handles user data and selections.
- [Terms of Service](https://limato.app/terms): The legal agreement governing use of Limato.
- [Refund Policy](https://limato.app/refund): The conditions under which paid plans can be refunded.

## Contact

- Support: support@limato.app