Sentence-level heatmap
The score is one number. The heatmap shows you which sentences contributed to it, that's the actionable signal.
For students · Pre-submit check
If you wrote the essay, you want to know whether a detector will wrongly flag it. If you used AI for part of it, you want to know which parts read as AI so you can rewrite them in your own words.
There are two reasons students check an essay before submitting it. The first is honest: you wrote the whole thing yourself, but you've heard detectors misfire on formal academic prose, and you want to know whether yours is at risk. The second is also honest in its own way: you used an AI tool somewhere in the draft, for brainstorming, for a rough first pass, for smoothing a paragraph, and you want to make sure the final version reads as your writing.
This page is for both cases. The workflow is the same: paste or upload the essay, read the sentence-level heatmap, decide whether to rewrite anything.
A high AI-likelihood score does not mean you cheated. It means the sentence patterns in your essay match the patterns the detector associates with AI-generated text. Those patterns are things like low perplexity (the next word is too predictable), low burstiness (sentences are all similar length), and uniform vocabulary. Careful academic prose sometimes scores high for exactly these reasons. That is a known failure mode.
The evidence that matters is the heatmap, not the overall score. If the heatmap lights up on five specific sentences in a 40-sentence essay, that is useful. If it lights up evenly across the whole essay, the detector is likely responding to your natural style, not to AI.
If you're confident you wrote every word and the detector is flagging the essay anyway:
Your school's policy is the first thing to check (your syllabus, or the course site). If AI is allowed with attribution: add the attribution. If AI is allowed for drafting but not for final prose: the flagged sentences are where you rewrite. Our rewrite guide walks through how to do this without making the essay worse.
If AI is prohibited in your class: the right move is not to sneak it past a detector. The right move is to rewrite those sections in your own voice, or to talk to your teacher. Trying to defeat a detector is a losing game, detectors improve faster than obfuscation techniques do.
Your essay is stored for 30 days so you can re-check it, then deleted. We never use submitted essays to train the detector. Full policy on /privacy.
The score is one number. The heatmap shows you which sentences contributed to it, that's the actionable signal.
We have no relationship with your school's gradebook, LMS, or detector. Checking your essay here is private.
Anonymous users get 5 checks per day. A free account raises that to 20/day. No card required.