What readability measures
Most formulas estimate the education level needed to follow your prose, based on sentence length and word length. A higher grade level means longer sentences and bigger words; a lower one means shorter, simpler text. The score is a proxy for effort, not intelligence.
Different formulas weight things differently, but they all reward short sentences and common words. That is why the same paragraph can read as "grade 8" in one and "grade 10" in another — treat the score as a range, not a verdict.
The formulas you will actually meet
Readability scoring is not one algorithm; it is a family of formulas developed between the 1940s and today, and each has a reason it exists.
- Flesch Reading Ease (Rudolf Flesch, 1948) returns a 0–100 score, not a grade. The formula is 206.835 minus 1.015 × average sentence length minus 84.6 × average syllables per word. Its bands: 90–100 is very easy, about a 5th-grade level; 60–70 is plain English for most adult readers; 50–60 reads like high-school material; below 30 is college-graduate territory. Consumer-facing writing usually aims for the 60–70 band. Flesch tuned the constants on the McCall-Crabbs reading test passages — which is one reason the score behaves oddly on text unlike 1940s school prose, such as dialogue-heavy fiction or code documentation.
- Flesch-Kincaid Grade Level (developed with the U.S. Navy in 1975) converts the same inputs into a U.S. school grade: 0.39 × average sentence length plus 11.8 × average syllables per word, minus 15.59. The Navy adopted it for technical manuals — a document that must be readable by an 18-year-old sailor with a specific training level. Because it outputs "grade 9.2," people misread it as a precise measurement. It is a linear formula; grade 9.2 means "the formula landed there," not "a 9th grader tested on this."
- SMOG (G. Harry McLaughlin, 1969) counts polysyllabic words in sentence samples and is the health-care favorite, largely because validation work tied it to actual comprehension of health materials. The original protocol wants 30 sentences — 10 from the start, middle, and end of a document. If a patient-education leaflet says "SMOG 6," that is why it was chosen over Flesch.
- Gunning Fog (Robert Gunning, 1952) came from business and newspaper consulting; it scores 0.4 × (average sentence length + percentage of "complex" words, defined as three or more syllables). Gunning's target was workplace prose — he coined the whole "foggy writing" critique against corporate and government gobbledygook.
- Dale-Chall (Edgar Dale and Jeanne Chall, 1948) is the odd one out: instead of syllable counting, it checks words against a list of roughly 3,000 words familiar to most 4th graders. This matters because syllable formulas can be fooled — "copay" and "clinic" are short but specialist; Dale-Chall catches them, syllable counters do not.
Glint's readability analyzer computes these scores in your browser so you can compare them rather than trusting any single one.
Where the standards come from
The push for plain language is regulatory, not aesthetic. The U.S. Plain Writing Act of 2010 requires federal agencies to write public documents in "clear, concise, well-organized" prose, and the federal guidelines live at plainlanguage.gov. Health communication goes further: guidance followed by agencies like the CDC and promoted by medical bodies such as the AMA has long advised aiming patient materials at around a 6th-grade reading level, on the logic that health literacy is life-safety. The benchmark study behind much of this — the 2003 National Assessment of Adult Literacy — found only about 12 percent of U.S. adults scored "proficient" in health literacy. Meanwhile, insurance policies and privacy policies famously score at college and graduate levels.
Why "no signup" matters
You might paste a draft client email, a patient leaflet, an unpublished internal memo, or a legal notice before it is public. A no-signup, browser-based checker processes the text on the page and does not store it. No account means no copy of your writing left on a server — which matters more than it sounds, because drafts are the most sensitive text you will ever run through a tool: they contain sentences you deleted for a reason and numbers not yet approved for release.
When to check readability
- Plain-language mandates where a public audience must be able to act on the information — government notices, consent forms, benefit letters.
- Blog and marketing posts where you want a smooth, approachable read; most mainstream web writing lands naturally between grade 6 and 9.
- Internal docs so new team members understand them without help.
- Non-native audiences who need clearer, shorter sentences — shorter sentences help far beyond the grade-level math, because each sentence holds one idea at a time.
- Email and support where clarity reduces back-and-forth; every grade level you shed is a ticket you never receive.
- Patient and safety materials, where SMOG-style checking is close to an industry expectation.
How to lower the score
Shorten sentences first. Halving average sentence length usually drops the grade level faster than swapping big words. Cut jargon or define it on first use. Use the score as a revision guide, then read the passage aloud to confirm it still sounds like you.
Why sentence length dominates
Look at the Flesch-Kincaid formula again: 0.39 × average sentence length plus 11.8 × syllables per word, minus 15.59. Sentence length and word length both count, but they do not contribute equally in practice. The average English word is about 1.6 syllables and that number barely moves as you edit — vocabulary is sticky. Average sentence length, meanwhile, can swing from 28 words in bureaucratic first drafts to 14 words after one disciplined editing pass, and every point of sentence length moves the score.
Concretely: take a 30-word sentence with two subordinate clauses and split it into a 20 and a 10. You have not changed a single word, and the grade level falls — because the formulas, whatever their differences, all encode the same finding from reading research: comprehension drops sharply as sentences carry more coordinated clauses. The clauses are the fog; the formulas just meter it.
There is a second-order benefit the formulas do not even measure. Long sentences force the reader to hold unfinished meaning in working memory. Short sentences let meaning land and reset. That is why "read it aloud" remains the gold-standard check: your breath runs out where your readers' attention does.
A workflow that actually works
- Write the draft without looking at any score. Scoring mid-draft makes you edit your thinking.
- Run the checker. Note the score, but also skim for the mechanical symptoms: sentences over 25 words, paragraphs over 6 lines.
- Fix sentences first. Split, cut subordinate clauses, convert passive to active. This alone usually moves the score most of the way.
- Fix jargon second. Replace or define specialist terms on first use. Do this after sentences, because shorter sentences often reveal which jargon is actually load-bearing.
- Rescore. If it still reads two or more grades above target, the problem is probably structure, not surface — reorder the argument.
- Read the final version aloud once. If a sentence trips your tongue, it trips your reader.
A worked example
Hard: "The implementation of the aforementioned protocol necessitates the utilization of cross-functional alignment to facilitate optimal outcomes."
Easier: "To get the best results, teams need to work together using the new protocol."
Same idea, lower score, no lost meaning. The second is simply easier to read. What changed, mechanically: one 16-word sentence stayed one sentence, but every heavy noun ("implementation," "utilization," "facilitate") became a plain verb or verb phrase. This is the nominalization fix — turning inflated nouns back into verbs — and it lowers scores almost as reliably as splitting sentences does, because inflated nouns are usually five syllables of air.
The main levers, ranked by how much they move the needle:
- Shorter sentences — the dominant driver of every formula.
- Common words — smaller syllable counts, and Dale-Chall-friendly vocabulary.
- Cut or define jargon — controls difficulty where specialist terms are unavoidable.
- Long chains of clauses — the fastest way to raise the score; untangle them.
Free vs paid checkers
Free tools give you the core score and basic tips. Paid ones add sentence-level highlights and suggestions. For most writers, a free no-signup checker plus careful editing is enough.
The ecosystem, honestly mapped: Microsoft Word has shipped built-in Flesch-Kincaid for decades — most people never find it because it hides behind File → Options → Proofing → "Show readability statistics," and only appears after a spell check. The Hemingway Editor popularized the color-highlighted sentence approach and shows a grade level, though its grade is the app's own weighting rather than a named formula tied to a school standard. Grammarly reports readability as part of its suite. Yoast bakes Flesch checks into WordPress SEO plugins. What none of the account-based tools give you is a no-upload way to check a draft — which is exactly the slot a browser-based checker fills.
Common mistakes
Writing down on purpose is the trap. A low score is good only if it fits the audience — a journal for specialists should read harder than a public warning. The tool tells you where you landed; judgment tells you whether that is the right place. Also, do not chase a number so hard that you lose precision or sound childish.
The deeper trap is gaming the formula. A score measures surface complexity — sentence length and word length — not clarity. You can split a tangled 30-word sentence into three short ones and keep the tangle, now just spread across three sentences with the same missing subject. You can swap "utilize" for "use" without fixing the illogical order of the argument. The score falls; the reader is no better off. Professional editors use readability scores the way pilots use instruments: as a check against self-deception, never as the flight plan itself.
Two corollaries worth internalizing. First, a score can be low and the text still fails — vague, disorganized text often scores beautifully because short, common words are exactly what vague writing uses. Second, a high score is sometimes correct: a paper for immunologists, a legal brief, an API reference. The audience sets the target; the formula only tells you whether you hit it.
Who should use it (and who shouldn't)
Use a readability checker for public information, marketing, onboarding, and any text read by a broad audience. Skip it for specialist or technical writing where the reader expects dense material — there, clarity still matters, but a low grade level is not the goal, and forcing one will strip the precision the reader came for.
The test is not "who is smart enough" but "who is motivated enough." Specialists will forgive dense prose because they need the content. The public will not forgive dense prose because they can leave.
How it fits a writing toolkit
Glint's readability checker is free and needs no account. Use it with the word counter: the counter tells you how much you wrote, the analyzer tells you how hard it is to read. Together they catch both "too short" and "too dense."
For deeper work, we have guides on how to reduce reading grade level with specific sentence surgery, and how to improve your Flesch Reading Ease score without dumbing the text down. If you are comparing dedicated tools, see our roundup of the best reading ease analyzer tools.
A quick scenario: fix a public notice
A city office posts a notice about water maintenance. The first draft reads at a grade 14 level — fine for the lawyers who wrote it, useless for many residents who need to act on it.
They run the checker, see the grade, and rewrite. "Pursuant to municipal code, remediation will commence" becomes "We will start repairs." "Residents are advised to secure an adequate potable water reserve" becomes "Please store some drinking water." The grade drops to about 7, and the notice finally tells people what to do.
Notice what the rewrite actually did, because it is the pattern behind every plain-language fix. The first draft was written to protect the writer — every hedge and citation of authority exists so nobody can blame the author. The rewrite was written to arm the reader — verbs, actions, ownership. The grade-level drop was a side effect of that shift in allegiance, not the goal itself. If they had chased the number directly, they would have produced short sentences that still failed to say "store drinking water."
That is the real value of readability checking: it catches the gap between the writer's world and the reader's. Public information exists to be used, and a notice nobody understands fails its only job. The checker makes that failure visible before publish, not after complaints.
The same applies to product updates, health guidance, and internal comms. Any text meant to prompt action should be checked for the level its readers actually have. Write to the reader, confirm with the score, and read it aloud once more. Clarity is a choice you can measure.
Frequently asked questions
Is a no-signup readability checker safe for private drafts? A browser-based checker processes text on the page and does not upload it, so your draft stays on your device. No account means no stored copy.
What is a good readability score? It depends on audience. Public information often targets around grade 6 to 8 — roughly the 60–70 band on Flesch Reading Ease — while specialist material can be higher. Match the score to the reader.
How do I lower a high grade level? Shorten sentences first, then cut or define jargon. Sentence length drives the score more than word choice because average word length barely changes in editing, while sentence length can be halved.
Should I always aim for the lowest score? No. Write for your reader. A low score is only good when the audience needs simple text — and a low score can hide disorganized writing.
Is the Glint readability checker really free with no account? Yes. It is free to use and requires no signup; the analysis happens in your browser.