Loading page
About this tool
Scan a social caption for AI-style hooks, em dashes, emoji bullet patterns and hashtag padding, with a transparent signal score.
The AI Social Caption Detector scores a social media caption on nine measurable stylistic tells found in unedited language-model output: em dashes, over-used vocabulary such as delve and tapestry, stock hooks like "stop scrolling", the "not just X, it's Y" construction, emoji bullet layouts, emoji density, hashtag padding, repeated rule-of-three lists and machine-even sentence rhythm. Each signal carries a fixed weight and the total is normalised to a 0-100 score you can see broken down line by line. It is a style check for editors and social managers, not an authorship classifier.
Open AI Social Caption Detector on AltFTool — it loads instantly in your browser.
Provide your input — an image, text, or data.
Let the tool analyze or generate the result.
Review, refine, and reuse the output wherever you need it.
Every point is attributed to a named signal with the exact matched text.
The caption is analysed in your browser, so client copy stays private.
Adds Instagram's 30-hashtag cap, LinkedIn hashtag guidance and the 280-character limit on X.
No. It measures style conventions, not provenance, and no tool can prove authorship from text alone. A human who writes in short emoji-led lines with em dashes will score high, and a carefully edited model draft will score low, so treat the result as an editing prompt rather than evidence.
Because the em dash (—) is not on a standard phone keyboard, so most people typing a caption on mobile use a hyphen, comma or full stop instead. Models insert em dashes freely, which makes a caption with several of them stand out stylistically — though desktop writers and anyone using autocorrect substitution use them legitimately too.
Instagram allows a maximum of 30 hashtags per post, and LinkedIn's creator guidance suggests around three relevant ones. This tool flags padding above 10 hashtags, or three or more generic reach tags such as #viral or #motivation, because those add bulk without describing the post.
It is the coefficient of variation of sentence lengths — the standard deviation of words per sentence divided by the mean. Below 0.35 the caption's sentences are unusually similar in length, which is a common trait of generated text; human writing normally mixes a three-word line with a twenty-word one.