What people want from a readability checker
People search for a readability checker at the moment a draft feels heavier than the idea. A teacher pastes a worksheet and wonders if the instructions are harder than the lesson. A founder rereads a customer update and cannot find the date. A student is told to "write at an 8th-grade level" and has no idea what that number measures. A support lead wants replies that a tired person can finish. The search is not for a literary theory. It is for a number that says, in plain terms, whether this paragraph is easy or hard.
The number people usually mean is one of two cousins. Flesch Reading Ease is a score that rises as the text gets easier, often talked about on a 0 to 100 scale. The Flesch-Kincaid grade level is a U.S. school grade estimated from the same ingredients: how long the sentences are, and how long the words are. A third formula, the Gunning Fog index, answers a similar grade-level question by counting "complex" words more directly. This page runs all three on the text you paste, then shows the sentences that did the most to push the score up.
A checker like this is a surface measure. It cannot tell a good idea from a bad one. It cannot tell a joke from a warning. It cannot know that your reader is a surgeon who wants the long word, or a new customer who needs the short one. What it can do, quickly and without sending the draft anywhere, is show the two levers you can actually pull: break a sentence, or trade a decorative long word for a short one. Most of the time, the sentence break does more.
Reading ease, grade level, and why both exist
Flesch Reading Ease was designed so that a higher score means easier reading. A score around 90 to 100 is very easy, the kind of line you might put on a sign. Around 80 to 90 is easy conversation. Around 70 to 80 is fairly easy, a good band for a mass-market explanation. Around 60 to 70 is plain English, roughly 8th or 9th grade, and it is the band many editors name when they say "write for a general adult." Around 50 to 60 starts to feel like high-school homework. Around 30 to 50 is college. Under 30 is the dense end, where even a careful reader slows down.
The Flesch-Kincaid grade level restates that difficulty as a grade. A 6 means about sixth grade. A 12 means about the end of high school. The grade can fall below 0 on a tiny, very easy line, and it can climb past 16 on a single overloaded sentence. Those extremes are the formula doing arithmetic, not a claim that the text is preschool or postgraduate in any official sense. There is no school board behind the decimal. There is an average.
People argue about which number to quote. Quote the one your reader asked for. If a style guide says "8th grade," report the Flesch-Kincaid grade. If a content brief says "reading ease above 60," report Reading Ease. If the two seem to disagree, look at sentence length before you look for a bug. Ease and grade are built from the same two averages, with different weights, so they move together. Fog can sit higher because it cares about the share of long words, not only the average syllable. A paragraph of short sentences sprinkled with necessary technical terms can look friendly on Reading Ease and stricter on Fog. That is useful information. It means the sentences are fine and the vocabulary is the part to review.
None of these scores measure comprehension directly. A person can score a text as "easy" and still misunderstand it, because the logic is wrong, the referent is unclear, or the reader lacks the background. A person can also read a "difficult" paragraph smoothly because they know the field. The score is a proxy that has held up well enough, for decades, to be the thing editors, teachers, and government plain-language guides still use. Use it as a proxy. Then read the paragraph out loud once. Your ear will catch the sentence the formula only marked.
The formulas, with a hand example
You do not need to do the arithmetic by hand. The checker does it. The formulas are short enough to trust, and a worked example keeps the buttons honest.
Flesch Reading Ease = 206.835 minus 1.015 times the average words per sentence, minus 84.6 times the average syllables per word. Flesch-Kincaid grade level = 0.39 times the average words per sentence, plus 11.8 times the average syllables per word, minus 15.59. Gunning Fog, as this page calculates it, = 0.4 times the sum of average words per sentence and the percentage of words that have three or more syllables.
Take a made-up paragraph that is exactly 100 words, 5 sentences, 140 syllables, and 8 words of three syllables or more. Average sentence length is 20 words. Average syllables per word is 1.4. Complex words are 8 percent. Reading Ease is 206.835 minus 20.3 minus 118.44, which is 68.1. That is plain English. The grade is 0.39 times 20, plus 11.8 times 1.4, minus 15.59, which is 8.7. About ninth grade. Fog is 0.4 times 28, which is 11.2. Same paragraph, three views: comfortable for a general adult, a bit high if you promised sixth grade, and a reminder that those 8 long words are doing some of the lifting. If you split the paragraph into 8 sentences and leave the words alone, the averages change and every score drops, even though you did not "dumb down" a single term. That is the practical lesson. Sentence breaks are the cheapest clarity you can buy.
The live tool estimates syllables instead of trusting a hand count. On a long paragraph the estimate is close enough that the grade usually stays in the same band. On a 15-word slogan, one missed syllable can move the decimal. The page says so when the paste is under 30 words. The plain sample and the dense sample are there so you can see a real swing without inventing a paragraph of your own. The plain sample is a launch note in short sentences. The dense sample is one sentence of committee language. The dense sample should mark itself as a long sentence. If it does not, the counter is wrong. If it does, you have seen the whole product.
Which band fits the thing you are writing
A voice message is the strictest everyday case, because the listener hears it once. Aim for a Reading Ease of 70 or higher, a grade around 7 or below, and no sentence over 20 to 25 words. A 20-second note does not have room for a clause that qualifies a clause. Say the change, the reason, and the date in three sentences. If a term of art is required, put it in its own short sentence and then say what it means. People forgive a plain voice note. They replay a tangled one, or they do not, and they ask you to type it.
Email and chat sit next to voice notes. A busy reader scans. A Reading Ease of 60 to 80 is a fair target for a customer, a classmate, or a teammate who is not already deep in the thread. The first sentence should carry the point, because many people never reach the second. A grade of 8 to 10 is not a failure in a professional email. A grade of 14 in a three-line scheduling note is a failure of editing, not of intelligence. Look at the marked sentence. It is usually the one that tried to include the context, the decision, and the apology at once.
Lessons and help articles can be a little denser, and they should still be kinder than the textbook they came from. A class handout around 60 to 75 lets a student spend attention on the idea. Instructions that score under 50 often fail for a boring reason: the step and the warning share a sentence. Split them. "Click Save. You cannot undo this." is easier than a 35-word sentence that mentions the button, the menu, the exception, and the consequence. The reading time calculator tells you whether the lesson fits the minutes you have. This checker tells you whether those minutes will be spent understanding or untangling.
Specialist writing has a different job. A methods section, a contract clause, or a clinical note among peers may land between 30 and 50 and still be the right text. The question is whether the difficulty is the subject or the costume. "Use the distal port" is short and still specialist. "It is incumbent upon the practitioner to utilize the distal port in order to facilitate administration" is the same instruction in a coat. The checker will score the second one as harder. It is harder. The term "distal" can stay. "Incumbent," "utilize," and "facilitate" can go. When you check specialist text, paste the audience-facing part separately from the technical part so one score does not average a clear warning with a dense citation.
Image for the readability guide: a tall stack of closely printed pages next to a single airy page and a pencil, showing the difference between a long sentence load and a short one.
Sentence length is the lever
Average words per sentence is the statistic worth watching even if you ignore the grade. Under 15 words, speech and signs feel light. Around 15 to 20, most adult prose feels normal. Around 20 to 25, readers are fine and listeners start waiting for the verb. Over 25, both groups work harder, and this page marks the sentence. The mark is not a punishment. It is a location. Open that sentence and find the joint. "And," "which," "that," "because," and a comma before an extra clause are the usual joints. Put a period there. Read both new sentences. If each one has a subject and a verb, you are done with that mark.
A long sentence is sometimes the right sentence. A list of three short conditions can live in one line if the structure is obvious: "Bring the receipt, the serial number, and a photo of the label." The checker may mark it if the list is long. Read it aloud. If you can say it without gasping or forgetting the first item, leave it. If you cannot, break the list into bullets or into two sentences. The formula does not understand bullets well if you paste them as one wrapped line without periods, because it may treat the whole list as a single sentence. Add a period to each step when you want a fair score, or accept that a list will look longer than it feels.
Short sentences can also go too far. Five words, then four, then three, starts to sound like a telegram or a threat. Vary the length once the monsters are gone. A 22-word sentence between two 8-word sentences gives the paragraph a pulse. The average can stay near 16 while the paragraph still sounds like a person. Chasing a grade of 3.0 by reducing every line to a stub removes the connective tissue readers use to see cause and effect. "The build failed. The key expired." is clear. "The build failed because the key expired." is also clear, and it is one word of logic the stubs left out. Prefer the version that keeps the because.
When you revise, change one kind of thing at a time. First split every marked sentence that you stumbled on when you read aloud. Check the score. If you are already in the band you wanted, stop. If the grade is still high, look at words of three syllables that are not the actual subject. "Additional," "currently," "individual," and "implementation" are frequent costume words. "Use," "now," "person," and "start" are not always legal substitutes, and they often are. Replace three costume words, not thirty. Then stop again. A second pass that hunts every remaining long word usually damages tone and precision together.
Syllables, long words, and the Fog index
Word length in these formulas means syllables, not letters. "Strength" is one syllable and looks long. "Idea" is short on the page and has three syllables. A character counter will not tell you the same story. That is why a readability checker and a character counter answer different questions, and why "business" has to be handled with care: the spelling hides a syllable. This page counts vowel groups, drops a silent final e, drops a silent -ed, and keeps a short list of common words whose spelling is a trap. The list includes people, business, created, science, every, family, and a handful of others. It is not a pronouncing dictionary.
The practical effect of a small syllable error is small on a real paragraph. The Flesch-Kincaid grade moves by about 0.12 for each hundredth of a syllable in the average. If the checker is off by 5 syllables in a 200-word note, the average moves by 0.025 and the grade moves by a fraction of a step. You will not cross from "plain" to "college" because of "business." You will cross because of a 40-word sentence. Trust the highlight more than the second decimal of the grade.
Fog uses a cutoff instead of an average. Any word of three or more estimated syllables counts as complex, and the percentage of those words is added to the sentence length. "However" counts. "Important" counts. "The" does not. Fog was built to flag fog, so it is stern. A grade of 12 on Fog does not mean you must rewrite until Fog says 8. It means about one word in this draft is carrying extra syllables, and you should know whether that word is "algorithm," which you may need, or "methodology," which you may not. Compare Fog with the Flesch-Kincaid grade. When they are close, the draft is consistently hard or consistently easy. When Fog is much higher, the sentences are reasonable and the vocabulary is the spike.
Proper names inflate every syllable score. A product named in three syllables, repeated ten times, is not fog. It is the subject. The same is true of a place, a medicine, or a person's name. If the score looks harsh and the highlights are mostly names plus one real monster sentence, fix the monster and leave the names. Abbreviations do the opposite. "URL" may be counted as one syllable or as letters, and either way it is shorter on the page than "address." Readers still have to know what it means. A checker will not punish an unexplained acronym. You should. Spell it once, in a short sentence, then use the short form.
Voice scripts, lessons, and the difference between reading and hearing
A paragraph that is pleasant to read can be unpleasant to hear. The eye can skip back. The ear cannot. That is why a voice message, a short lesson, and a podcast open deserve a stricter band than an article on the same topic. Take the plain sample on this page. It is a launch delay told in short sentences: what moved, what is missing, and what the note should say. Record something in that shape and a listener gets it. Take the dense sample, which is one sentence about a committee postponing a timeline until assets are finalized for stakeholder review. It may be accurate. It is not sayable. The highlight exists so you do not have to feel that difference vaguely.
Before you record, check three things in this order. First, readability, so the sentences fit a breath. Second, fillers, if you are pasting a transcript rather than a script, using the filler word counter. Raw speech often looks simpler than it is, because um and you know are short, or harder than it is, because restarts create fake sentences. Third, duration, using the speaking time calculator. An easy 250-word script is still too long for a one-minute slot. Ease does not refund seconds.
If the take is already recorded and the wording is the problem, a score will not rewrite the audio. It tells you whether a retake is about clarity. When the performance is good and the sentences are the mess, clean the voice message instead of chasing a grade on a transcript you will not send. VClar's grammar repair for voice messages works on the spoken wording, and the filler words remover cuts the verbal clutter while keeping the speaker's voice. Use this checker on the script or the cleaned transcript. Use those tools on the recording. They are different edits, and doing all of them in your head while you hold the phone is how the second take gets stiffer than the first.
Practice image for the readability checker: a person speaks a voice message into a phone in a bright room, with a short script turned face down so the words stay plain enough to say.
How to lower a score without hollowing the draft
Start with the longest marked sentence. Read it aloud and stop at the first place you wanted a breath. That place is usually a legal period. Write the two sentences. Do not add new facts while you split. A revision that sneaks in a second idea will create a new long sentence somewhere else. Check the score. One split on a short note can move the grade by a year. That is allowed. It is also a reason not to celebrate a perfect number on a 40-word paste until you have read it.
Next, hunt costume verbs. "Utilize" is often "use." "Facilitate" is often "help" or "run." "Prior to" is "before." "In order to" is "to." "Due to the fact that" is "because." These swaps lower the syllable average and, more importantly, they give the sentence back its verb. Leave the words that would change the claim if you shortened them. "Not" and "except" are short and essential. "Approximately" sometimes means you truly do not have an exact figure, and sometimes means you are nervous. If you have the figure, use the figure. Numbers are easy to read and specific, and a checker will not scold you for "Friday" or "4 p.m."
Then read the first sentence alone, the way a scanner or a listener will. If it does not contain the result, move the result forward. Readability formulas do not reward front-loading, and readers do. A paragraph can score 75 and still fail because the date is in sentence four. The checker will not move that date for you. After the score is in band, spend one minute on order. Result, reason, ask. The same order works for a voice message, a support reply, and the opening of a lesson.
Stop when a reader in the real audience would not stumble. A further drop from 8th grade to 4th grade is not free. It costs examples, hedges that are actually true, and the rhythm that makes a paragraph pleasant. If you are writing for children, keep going. If you are writing for adults who asked for plain language, 7th to 9th grade with no 30-word sentences is a successful edit. Publish that. Another hour of syllable shaving produces text that sounds like it was written to satisfy a formula, because it was.
Teachers, students, support teams, and founders
Teachers can use the checker on the instructions, not on the poem. A poem's grade is meaningless. A lab sheet's grade is not. Paste the directions the student must follow before the experiment starts. If the steps score as harder than the concept you are teaching, the sheet is the lesson and the concept never gets a turn. Split the steps until a student can read each one without help. Leave the scientific term in, and define it in the next short sentence. Students who are told "write clearly" finally have a target: bring the narrative portion to a band you name, such as Reading Ease 60 or better, and do not apply that target to a quotation or a formula.
Students pasting their own essays should watch for a different trap. Padding raises the grade in the bad direction. A sentence that exists to sound formal, full of "furthermore" and "it is important to note that," is both longer and less clear. Cut the throat-clearing and the grade often falls into the range the assignment wanted, while the argument gets shorter and stronger. If the assignment is a research summary for classmates, 9th to 11th grade is a respectful target. If it is a reflection, you can be easier. Do not chase a 6 if the rubric wants analysis. Analysis can live in short sentences. "The result contradicts the hypothesis because the sample was too small." is analysis, and it is not a hard sentence.
Support teams benefit from a shared band more than from a shared adjective. "Keep it readable" means something different to every new hire. "Reading Ease 70 or above on the reply the customer sees, and no sentence over 25 words" is a rule you can paste into a macro and check. Internal troubleshooting notes can be denser. The customer-facing paragraph should not inherit the density of the log. Paste only the paragraph you will send. Founders have the same split. The changelog can name the subsystem. The voice note to customers should say what they can do now, in sentences a person can repeat after one hearing. If you cannot repeat your own note, the checker is about to mark a sentence, and the customer is about to ask you to send an email.
What the score gets wrong
English is the assumption. The weights were fit to English sentence rhythm and English syllable patterns. They are the wrong instrument for a French, Hindi, Japanese, or Spanish original. Sentence length still gives you a hint in those languages. The grade number does not. If you are preparing an English version of a message to send or to record, score the English. If you are preparing a message that will stay in another language, do not cite this grade as that language's reading level. VClar's language guides cover voice messages across the public languages on the site. This checker stays an English text tool.
Dialogue, poetry, and fragments confuse sentence counts. A line break is not a period. If you paste a script and forget the punctuation, the checker may see one enormous sentence and report a frightening grade. Add the periods and question marks the speakers actually used, then score again. Headings without periods can stick to the following sentence. That usually makes the text look slightly harder. Put a period on a heading if you want it counted as its own short line, or leave it and know the average moved a little.
Quotations should be scored with the same honesty you would use in a review. If you are quoting a statute, the grade of the quotation is the statute's grade, not your failure. Paste your explanation separately. If your explanation is harder than the quote, the explanation is the problem. The checker has no idea which words are yours. Parentheses and dashes are also invisible to the "long sentence" rule except as part of the word count. A 24-word sentence with a 15-word parenthesis may pass the 25-word mark and still be hard. Read those aloud. The rule is a net, not a ceiling on difficulty.
Very short texts flicker. Under 30 words the verdict on this page tells you the grade can swing. A headline is not a paragraph, and running Flesch-Kincaid on four words produces a novelty number. Wait until you have several sentences. At the other end, a 5,000-word chapter can hide three unreadable sections inside a friendly average. If the chapter matters, score the introduction, the warning, and the conclusion on their own. Averages are kind to documents that are uneven. Readers are not. They remember the paragraph that lost them.
The checker also cannot see layout. A wall of easy sentences with no paragraph break still feels hard. White space, headings, and a visible next step do not change Reading Ease and they change comprehension. After the score is in band, look at the page as a page. One idea per paragraph. The action before the background, unless the audience truly needs the background first. Those choices are editorial. They are also the reason a perfect 72 can still be a bad help article, and a 58 with a clear structure can still be a good one.
A small set of drafts, read as numbers
A 45-word customer note in three sentences, with ordinary words, often lands near a Reading Ease of 75 and a grade around 6 or 7. That is a sendable note. If the same facts are crushed into one sentence, the ease can fall into the 40s and the grade can jump several years, even though the vocabulary did not change. The repair is a period, not a thesaurus. This is the pattern behind most "my update sounds stiff" complaints. The writer was not using rare words. The writer was using one sentence.
A 180-word lesson intro with an ease around 65 and two marked sentences is a normal first draft from an expert. The expert wrote the way they talk to colleagues. Split the two marked sentences and the intro usually enters the 70s without losing the term the students came to learn. If you also replace every technical term, the score rises again and the lesson becomes vague. Keep the term. Define it once. Move on. Students can hold a definition that has its own sentence. They struggle to unpack a definition that is a clause inside a clause.
A 600-word help article at ease 55, Fog 13, and six long sentences needs structural editing, not a new voice. Attack the six sentences. Then score the "If this failed" section by itself. Troubleshooting sections collect conditions, and conditions collect length. "If the save fails, check the connection. Then try again. If it fails a second time, copy the error code." will beat one sentence that contains all three moves. The article's overall grade will follow the sections you fixed. You do not need every example to score like a billboard.
A legal or policy paragraph at ease 25 can be acceptable and still contain one sentence the customer must not miss, such as a fee or a deadline. Pull that sentence out, score it alone, and rewrite only that sentence until it is plain. Leave the rest of the clause for counsel. A single blended grade would have told you the whole policy was "very difficult," which you already knew, and would have hidden the one line that decided whether the customer was surprised later. Separate scores are more honest than a heroic average.
Privacy and how to use the number tomorrow
The draft never has to leave the browser for these formulas to run. There is no login and no file upload on this page. The highlights are built on your device. If the text is a patient note, a student record, a personnel message, or an unreleased launch, treat the screen accordingly: do not leave it open, and clear the box when you stand up. Browser extensions and screenshots are outside what the page can promise. Close those if the draft is sensitive, the same way you would in any other editor.
Keep the number with the draft if you are comparing versions. "Ease 61, two long sentences" next to Monday's paste and "Ease 74, none" next to Tuesday's paste is a record of an edit, not a moral score. Use the same tool each time so the syllable estimate does not change underneath you. Bring a voice script back here before you time it. Bring a customer reply back here before you save it as a macro. Ignore the checker for poetry, brand lines of three words, and any sentence whose difficulty is the point, such as a quotation you are about to explain.
When the band matches the audience and the long sentences are either gone or deliberate, you are finished with readability. The next questions are different. How long will it take to read, how long will it take to say, and does the spoken version still carry filler. Those live on the other tools. This one has a narrow job, and it is the job behind the search: tell me, with a known formula, whether this English is easy or hard, and show me the sentence that made it hard.
Questions about the readability checker
What is a readability checker?
A readability checker estimates how hard a piece of English is to read. This one uses three published formulas on the text you paste: Flesch Reading Ease, the Flesch-Kincaid grade level, and the Gunning Fog index. It also counts words, sentences, and estimated syllables, and it marks any sentence of 25 words or more. The scores describe the surface of the writing, sentence length and word length. They do not judge whether the facts are true, whether the tone is kind, or whether a specialist reader needs the long words.
What is a good Flesch Reading Ease score?
For a general adult audience, 60 to 70 is the usual plain-English band, about 8th or 9th grade. A client update, a support reply, or a class intro is often clearer at 70 to 80. A voice message that someone will hear once, with no chance to reread, is safer at 70 or above, with no sentence over 25 words. A score under 50 means college-level density. That can be the right score for a paper written for researchers. It is a hard score for a 30-second voice note. Above 90 is very easy, which helps a short instruction and can sound thin in a technical discussion.
What does the Flesch-Kincaid grade level mean?
The grade level is an estimate of the U.S. school grade that would find the text comfortable. A 7.2 means about seventh grade, not that a sixth grader cannot read it and not that a manager should be embarrassed by it. The number comes from average sentence length and average syllables per word. Short sentences and short words pull the grade down. One 40-word sentence can lift the grade of a short note by a full year. Treat the number as a draft signal. Read the marked sentences, then decide which long words are doing real work.
Is a lower grade always better?
No. A lower grade means shorter sentences and shorter words. That helps when the reader or listener is in a hurry, new to the topic, or hearing the words once. It hurts when you delete a precise term and replace it with a vague one. "Invoice" should stay "invoice." "Myocardial infarction" can stay in a note between clinicians and should usually become "heart attack" in a note to a family. Match the grade to the audience. A terms-of-service clause that scores like a picture book has probably dropped a condition that mattered. A launch update that scores like a statute has probably buried the date.
How is this different from a reading time calculator?
Reading time answers how long the text takes. Reading level answers how hard each sentence is. A 200-word note can take one minute to read and still be college-level if the sentences are long and the words are long. A 2,000-word lesson can be easy if the sentences stay short. Use the words to time calculator when you need minutes. Use this readability checker when you need to know whether those minutes will feel smooth. If the text will be spoken, also check duration in the speaking time calculator, because silent reading time is shorter than speaking time.
Does the checker work on a voice message transcript?
Yes, with one caution. A transcript full of um, you know, and restarts looks harder than the idea underneath it, and sometimes easier, because the fillers are short words. Run the filler word counter first if the paste is a raw transcript, then check readability on the version you would actually send. A voice message should usually land at a Flesch Reading Ease of 70 or higher. Listeners cannot glance back at the previous line. If the checker marks a sentence, that sentence is the one to split before you record.
Why did the grade jump when I changed one sentence?
Short texts are unstable. The formulas divide by the number of sentences and the number of words. On a 40-word note, splitting one sentence in half changes the average a lot, so the grade can move by a year or more. That move is real for that note, and it is also why you should not compare a 40-word sample with a 900-word lesson at the level of a tenth of a grade. Once you are past about 100 words and five sentences, the score settles. Under 30 words, this page warns you that the grade can swing.
How accurate is the syllable count?
It is an estimate from spelling, not a dictionary. "Cat" is one syllable. "Table" is two. A few common words are listed by hand because English spelling hides syllables, including people, business, created, and science. Unlisted words are counted from vowel groups, with a silent final e and a silent -ed removed when that is the usual pattern. Names, brand terms, and words from other languages can be off by a syllable. A one-syllable miss on a long text barely moves the grade. On a single sentence it can. If a score looks wrong, read the long-sentence list. The sentence length is exact even when a syllable is estimated.
What is the Gunning Fog index?
The Gunning Fog index is another grade-level estimate. It uses average sentence length plus the percentage of complex words, and this checker treats a complex word as one with three or more estimated syllables. Fog often sits a little higher than Flesch-Kincaid on the same paste. Use it as a second opinion, not as a competing verdict. If ease says "plain English" and Fog says grade 11, you probably have a few long words inside otherwise short sentences. Keep the long words that are the actual subject, and cut the ones that are only formal dress.
Can I check languages other than English?
The formulas were built for English. Sentence length still means something in many languages. Syllables per word do not transfer. A French or German paragraph scored with these weights will produce a number that looks like a grade and does not describe that language's reading level. Paste English, including an English translation you are about to send. If you are checking a voice message you recorded in another language, score the English wording listeners will read or the English script you will record, not a machine translation you do not plan to use.
Does the readability checker store what I paste?
No. The counts run in this browser tab. There is no account on this page and no upload. VClar does not receive the text. Refreshing the page clears it. Copy anything you want to keep before you close the tab, and clear the box before you leave a shared computer. The score is only as private as the screen you are looking at.
Should legal, medical, or scientific text score at a 5th-grade level?
Only the parts a general reader has to act on. A consent step, a dose, a deadline, and a fee should be easy to read. The underlying method, citation, or statute can stay precise for the audience that needs it. Forcing an entire paper down to 5th grade often deletes the term a specialist uses to avoid a dangerous ambiguity. Split the document in your head: the action in plain sentences, the detail in the terms of art. Check those sections separately. One blended score hides the sentence the patient or the customer actually needed.
The readability checker is a free VClar tool for English text. When the words will be spoken, pair it with the speaking time calculator so the clearer draft still fits the time you have.