p3n

AI Detection

Paste text and get the probability that it was written by a model, with the sentences that give it away.

What you get

  • score: 0 (human) to 1 (AI).
  • verdict: likely_human, mixed or likely_ai.
  • flagged: the sentences that carry AI tells, and which tells.

How it works

p3n combines model-based classifiers with checks for the mechanical habits language models share. A score is a probability, not proof. Treat it as one piece of evidence.

Preview

This is a preview. The full detector, combining several models with sentence-level scoring, ships soon and keeps the same API.