Skip to content
EN
English 简体中文 soon 日本語 soon

IsGPT AI

Privacy-oriented AI detector

Visit official site

What IsGPT is

IsGPT is a detector built on the familiar statistical signals: perplexity and burstiness. It analyses submitted text to estimate whether a model produced it, covering academic, business, creative and technical material, and presents itself as privacy-friendly for institutional use.

What you can do with it

  • Check documents for AI signals
  • Screen papers, reports and proposals
  • Work with multilingual text
  • Review results inside an institution
  • Keep checks ad hoc rather than routine

Who it is for

  • Educators and institutions
  • Editors checking submissions
  • Researchers verifying material

What to watch out for

  • Perplexity-based detection has the same failure modes as its peers: formulaic or non-native human writing is flagged, edited AI text often passes
  • Institutional association does not make a statistical estimate reliable; treat results as signals only
  • Accusations require process evidence and a conversation, never a score
  • Confirm what happens to submitted text and whether it is retained

Pros & cons

✓ What we like

  • Clear, familiar methodology
  • Privacy-conscious framing
  • Free to use

! What to watch out for

  • Same reliability limits as other detectors
  • Affiliation is not validation
  • Not sufficient for any accusation

FAQ

Does its research origin make results reliable?

No. The methodology still misclassifies human writing.

Can it be used as evidence?

Only as one input alongside process evidence and discussion.

What should I check?

Whether submitted text is stored and for how long.

Last reviewed: 2026-09-19

More AI text detection tools

View all →

How we review