FREE TOOL
Does your LinkedIn post sound like AI?
Paste a post. The phrases that give away a machine-written draft are highlighted, each with what it does to a reader and what to do instead.
Checked in your browser against 93 phrases. Nothing you paste is sent anywhere.
What this checks, and what it doesn't
It checks for phrases that language models use far more often than people do: lines that point at the point (“that’s the part that matters”), stock shapes (“it’s not X, it’s Y”), stock openers and sign-offs, and a short list of words models overuse. The list comes from published research on AI writing and from drafts we have read ourselves.
It is not an AI detector, and nothing reliably is. A post with none of these phrases can still sound generic, and a person can say any of them. What it does well is catch the lines your network has learned to skim past.
Why these phrases cost you
People on LinkedIn read a lot of generated posts now, and they recognise the shapes quickly. One stock phrase in the first two lines and a reader decides the rest is generated, whatever is underneath it. The fix is rarely a better synonym. It is saying the specific thing the phrase was standing in for: what happened, what it cost, what you believe.
How On Record avoids them
On Record drafts from an interview with you, so the material is yours: your stories, your numbers, the way you phrase things. The same list is in its drafting instructions, and a draft that still contains one of these phrases is flagged to you before you post, unless it is something you said yourself in the interview.
Start with the interview
On Record interviews you about your own work, learns how you talk, and drafts posts from what you said. The first interview is free and needs no card.
Start free