Was that post written by a human? Check any social link.
Paste a link to a public Instagram, X, TikTok, Threads, Reddit, LinkedIn, Facebook or YouTube post. We pull the caption and tell you how much of it reads like AI-generated text — with the exact sentences that triggered the score — and read whatever provenance the post image still carries.
Works with public posts on Instagram, X (Twitter), TikTok, Threads, Reddit, LinkedIn, Facebook and YouTube.
How it works
Paste the link
Copy the post URL straight from the share menu or your browser's address bar. The post has to be public — we cannot read private or deleted posts.
We pull the caption
We read the post's public caption the same way a link preview does. No login, no scraping of your account, and we never ask for your social credentials.
You get a sentence-level verdict
The caption is scored for AI-generated writing patterns, and every sentence that pushed the score up is listed so you can judge for yourself.
What this tool can and cannot tell you
What it does well
- +Flags captions written by ChatGPT, Claude, Gemini and other large language models, with the specific sentences that look generated.
- +Works on longer captions — the kind used by brands, marketers, engagement-farming accounts and recruiters on LinkedIn.
- +Gives you the reasoning, not just a number, so you can disagree with it.
Where it stops
- –We read the image's provenance metadata (Content Credentials, generator fingerprints), but we do not judge the picture from its pixels. In practice social platforms strip that metadata on upload, so the image check usually returns 'no data' — which is an honest unknown, not a clean bill of health.
- –Very short captions cannot be judged reliably. Under 20 words we refuse to guess rather than invent a confident score.
- –No AI detector is perfect. A high score means the writing has AI-like patterns, not proof — treat it as a signal, not a verdict.
- –Private, deleted, or age-restricted posts cannot be read. Some platforms also block automated readers; when that happens, paste the caption text into our text detector instead.
Supported platforms
Paste any public post link from these networks.
Frequently asked questions
Do I need to log in to Instagram or X?
No. We only read what is already public on the post — the same information a link preview shows when you paste a URL into a chat app. We never ask for your social media password and we cannot see private accounts.
Can you tell if the photo or video is AI-generated?
Partly, and only when the file lets us. We read the image's provenance metadata — C2PA Content Credentials, the IPTC source type, and generator fingerprints. The catch is that Instagram, X and the rest strip all of that when a post is uploaded, so for most posts the answer is 'no data available'. We say that plainly instead of guessing from the pixels. If you can get the original file, our AI image checker will read far more from it.
Why does it refuse to score very short captions?
AI detection works by measuring patterns across sentences. Under roughly 20 words there is not enough signal, and any score would be close to a coin flip. Rather than show a confident-looking number we cannot stand behind, we tell you the caption is too short.
A post I know was human-written got a high score. Why?
Some people naturally write in the clean, structured, evenly-paced style that language models produce — and marketing copy in particular has always sounded a bit like this. That is why we show the flagged sentences and the reasoning instead of only a percentage. If the reasoning does not convince you, trust your judgment.
Is it free?
You get free checks each day without an account. A free account raises the limit, and a paid plan unlocks the full sentence-level report on every post plus the rest of the Plagly toolset.
Checking more than social posts?
The same detection engine works on documents, essays, and whole web pages.
Try the full AI detector