Paste the post into the box above and press Humanize. It rewrites every sentence in one pass at one credit per fifty words, so a 200-word post costs four of the ten credits you get free every day. No account, no card, no email address to verify. That handles the sentences. The other half of the job is the shape of the post, and on LinkedIn the shape is the loudest tell you have. The hook, the ladder of one-line paragraphs and the closing question are visible to every person scrolling past, and a rewriter works on sentences rather than on structure. This page covers both halves.
Your reader has already scrolled past a few hundred posts built from the same parts this month. They do not need software to recognise the next one. They recognise it by the second line and keep going. That is the whole penalty, and it arrives quietly, which is why it is easy to keep paying.
The template has five parts, and an AI-written post usually ships with all five:
Any one of these is a stylistic choice. All five together is a genre, and the genre now reads as generated even when a person typed every word.
AI prose reaches for a narrow set of constructions. On this feed the set is narrower still, because the model learned the genre from posts that were already imitating each other. The reference page at AI writing tells explains the general cases one by one, and the AI vocabulary checker marks the stock words in your own draft, free and unlimited, in your browser. Below are the phrases native to LinkedIn.
Delete each one and read the post again. It almost always gets shorter without saying less, which tells you what the phrase was there for.
Here is a draft in the house style. Every sentence sat on its own line with a blank line under it, which is how it lands on the feed. Nothing is wrong with the writing. You have simply read it before.
"I turned down a $50k client last week. Everyone told me I was crazy. But here's the thing. The best growth doesn't come from saying yes. It comes from saying no. Let that sink in. 3 lessons I learned: 1. Protect your energy. 2. Clarity beats capacity. 3. Your calendar is your strategy. Agree? What's the hardest no you've ever said? #Leadership #Growth #Entrepreneurship"
Nine lines and not one fact. No client, no reason, no consequence. The three lessons would sit just as comfortably under a post about hiring. Here is the same story with the story left in:
"I turned down a $50,000 client last week. They wanted weekly calls at 7am my time, a scope that had already moved twice before anything was signed, and a start date nine days out. I said no on the second call and felt sick about it all afternoon. Two weeks later the same money arrived from a project I could actually staff. That was luck, and it could not have happened if the calendar had been full. If you are weighing something similar this week, the fee is the easy part to compare. Compare the nine days."
The ladder of one-line paragraphs began as a real tactic. The feed cuts a post off after a few lines on a phone, so writers put a paragraph break where a comma belonged to buy more space above the fold. A model trained on those posts copies the white space without the reason for it, and you get a break every seven words in a post that was never going to be truncated.
Write in paragraphs of two to four sentences. Keep one standalone line if you have earned it, and only one, because the second one cancels the first. Cut the emoji bullets, which read as a slide deck rather than as a person. Cut the hashtag block to two tags or to none.
The closing question deserves one thought before you delete it. A real question aimed at specific people does get answers. "If you have run this on a team of more than fifty, I want to know whether the second rule held" is a question. "Thoughts?" is a request for engagement addressed to nobody.
The humanizer takes 12,000 characters per run, about 2,000 words, so a post never comes close to the ceiling. It costs one credit per fifty words with a 25-word minimum, which puts a 150-word post at three credits and a 250-word post at five. The free tier is ten credits a day per network with nothing to sign up for. That covers two 150-word posts rewritten and checked, with two credits left over for a short third rewrite.
A check costs one credit per run whatever the length and needs at least 65 words. For a post below that floor, check the longer draft you trimmed it down from, then trim the rewrite by hand. A check returns a 0 to 100 score and one of three verdicts: reads as AI-generated, borderline, or reads as human-written.
The rest of the toolkit is free and unlimited, and on a post-length draft it does most of the work anyway: the AI vocabulary checker, a proofreader, a readability checker, a plagiarism and originality check, a word counter and a signed drafting record. All of them run in your browser. The word counter reports characters as well as words, which is the unit a post box counts in.
Check the draft, then check the rewrite, and read the two numbers side by side. That is an outside read on the prose before the feed gives you one.
The score measures how machine-made the writing reads. It does not identify an author, because the writing is the only evidence in the box. Careful, formal, evenly paced prose collects the same flags from any scoring tool whether or not a model was involved. Liang and colleagues at Stanford HAI found in 2023 that 61% of TOEFL essays written by non-native English speakers were falsely flagged as AI-generated across seven detectors, which matters if you write in English as a second language and someone makes that accusation in a comment thread.
If you post publicly and expect that argument eventually, the signed drafting record logs your session as you write and exports a file, signed with ECDSA P-256, that anyone can verify without contacting us. Free, unlimited, no account.
One line on the rules. Rewriting a draft you were permitted to generate is ordinary editing. Where AI is not permitted, that rule covers the draft however it is edited afterwards. On LinkedIn the policy that binds you is your employer's or your client's, so read that one.
We make no claim about what any other detector will output. What happens here is concrete: every sentence gets rewritten in one pass, the AI vocabulary checker marks the stock words where they sit, and you get a 0 to 100 score with a verdict before and after. On LinkedIn the audience matters more than any classifier anyway, and the audience is reading for the hook, the ladder and the closing question.
Ten credits a day per network. The rewrite costs one credit per fifty words, so a 150-word post is three credits and a 250-word post is five. A check is one credit per run at any length. Ten credits covers two 150-word posts rewritten and checked, with two credits left for a short third rewrite, and there is no account and no card involved.
The checker needs 65 words, so a 40-word post is below the floor. The humanizer will take it, since its minimum is 25 words. For anything shorter than 65 words, check the longer draft you cut it down from, or use the AI vocabulary checker and the proofreader, which run in your browser with no minimum and no daily limit.
No, if it is a real sentence. "I turned down a $50,000 client last week" is a hook and also a fact. Delete it when it is engineered rather than said: a fragment with no subject, a claim with no number, or anything whose only job is to earn a tap on see more. The test is whether the line would survive in the middle of the post.
Because the genre trained the humans first. The one-sentence paragraph, the three lessons, the negation pivot and the closing question were LinkedIn conventions before models learned them. A person writing to the convention produces text with the same low variation in sentence length and the same stock phrasing, and that is what a score responds to.
It depends whose rules you are under. Rewriting a draft you were permitted to generate is ordinary editing. Where AI is not permitted, that rule covers the draft however it is edited. On a professional feed the relevant policy is usually your employer's or your client's on disclosing AI-assisted content, so read that one.