01

Input & Tokenization

The process begins when you paste your text or upload a document. Our system instantly breaks down the content into “tokens” (words, sub-words, and punctuation). This tokenization is similar to how AI models like GPT process information before generating a response.

02

Pattern Recognition

Our advanced algorithms analyze the text for specific linguistic signatures. AI models tend to write in very predictable patterns. We calculate the probability of each word following the previous one. If a text consistently chooses the most “probable” next word, it is likely AI-generated.

03

Perplexity & Burstiness Check

We measure two key metrics: Perplexity (how complex the text is) and Burstiness (the variation in sentence structure). Human writing is “bursty” — it has a mix of short and long, simple and complex sentences. AI writing tends to be flat and uniform.

The Science Behind It

N-Gram Analysis

We analyze sequences of N items (words) to detect statistical predictability common in Large Language Models (LLMs).

Burstiness

Humans write with varying sentence lengths and structures. We map these fluctuations to distinguish human “chaos” from AI “order.”

Cross-Verification

Our system runs your text against signatures from multiple models to ensure broad coverage.

Ready to Test It Out?

See the technology in action. It takes less than 5 seconds to verify your content.

Start Scanning