AI search manipulation safety checker

Flag categories of model-directed, concealed or misleading persuasion while keeping legitimate content work separate.

The text is examined only in this browser and is not uploaded, stored or logged.

Pattern matching can miss attacks or flag benign text; it is not a security certification.

Method

Flag categories of model-directed, concealed or misleading persuasion while keeping legitimate content work separate.

Privacy

The text is examined only in this browser and is not uploaded, stored or logged.

Limitations

Pattern matching can miss attacks or flag benign text; it is not a security certification.

Primary sources