Free, no account · method 0.8.0

Find passages worth checking before you publish.

Review generic wording, claims without visible evidence and repetition. Each signal points to a concrete editorial action.

  • Your text stays on this device
  • French and English
  • No AI origin score

French and English methods are experimental. Their real-world accuracy still needs independent evaluation.

Generic wording

Open with a fact your reader recognizes.

Review a text

The analysis runs in your browser. Your text is not sent.

Format changes advice, not detected passages.

0 / 12,000120 to 12,000 characters

Experimental English method: highlighted passages require human review.

Visible reasons to review.

See the exact passage

Each signal quotes the wording that needs your attention.

Choose an action

Check a source, remove a repetition or make a claim more specific.

Keep the limits in view

No AI origin score, accusation or invented certainty.

Misreadings, corrected

Public claims that distort a study or a number, with the correction, the source and its limit. These cases show the checking method. They do not measure the tool’s accuracy.

Loading published cases…

What this review does and does not do.

Does this detect AI-generated text?

No. It highlights passages for review and cannot establish who or what wrote them.

Why “AI slop” if the tool doesn’t detect AI?

Because that’s the name everyone uses. Slop is first a human way of writing, hollow and generic, long before chatbots. AI made it free and everywhere. The Observatory judges a passage, whoever wrote it, to help you write like yourself rather than like everyone else.

Does a clean result prove the text is accurate?

No. You still need to check facts, sources, context and usefulness.

Is my text uploaded?

No. Analysis runs in your browser. Your text leaves it only if you click “Rewrite in my voice”: it then opens in Kaneme, another YIRIWA, LLC service. Your draft may be held briefly in this tab while switching interface languages, then removed after restoration.

Can I analyze English text?

Yes, in experimental mode. Select the language of your text. Each language still needs an independent annotated evaluation.

Why is there no single score?

A score would conceal differences between evidence, repetition and editorial value.

Is every AI-assisted text slop?

No. Origin and quality are separate questions. Review the text’s value, sources and verification effort in its own context.

What happens to a submitted case?

It enters a private review queue. Nothing is published automatically.

What does the site count?

Six actions per day and language (page view, example loaded, analysis started, analysis completed, error, feedback sent), with no text, cookie or identifier. You can opt out from the privacy policy.

Review your next text before you publish it.

Review a text