Scan your writing before Crossplag does
See exactly what Crossplag would flag in your writing. Per-sentence detection scores with targeted suggestions to fix false positives.
About Crossplag's AI Detection
Crossplag uses perplexity and burstiness analysis to detect AI content. It looks for statistically smooth text with predictable vocabulary choices.
GPTypo scans for the same patterns, showing you sentence-level risk so you can fix flagged areas before submission.
What Crossplag looks for
These are the types of signals that Crossplag's classifier measures. GPTypo checks for the same patterns.
Sentence uniformity
AI text tends to produce sentences of similar length and rhythm. High variance is a signal of human writing.
Low perplexity
Predictable word choices with few surprises. Human writers are messier and less statistically smooth.
AI vocabulary
Overuse of phrases like "furthermore," "it is important to note," and "multifaceted" that appear far more often in AI output.
Structural repetition
Repeated paragraph openers, parallel sentence structures, and formulaic transitions between ideas.
Missing texture
Human writing has contractions, hedging, asides, and first-person markers. AI text tends to be uniformly polished.
Document coherence
AI text often has a narrow vocabulary range and smooth sentence-to-sentence flow that reads as artificially consistent.