content-scanning

To detect hate speech and spam in user-generated content, you can scan text or mixed media to identify policy violations. The system checks for multiple categories like violence and scams, providing per-category confidence scores and a PASS, REVIEW, or REMOVE recommendation based on your configured thresholds.

Yes, you can perform pre-publication checks to scan content for PII exposure. The scanning process identifies personally identifiable information alongside other violations like misinformation and copyright issues, flagging exact excerpts with precise locations and timestamps so unsafe content is caught before it goes live.

The best way to moderate bulk user content for policy compliance is to run automated bulk scans across your text and mixed-content sources. This process evaluates posts against multiple categories like hate speech and spam, producing structured results with flagged excerpts and actionable recommendations to streamline your moderation workflow.

Yes, content scanning works for triaging reported posts by evaluating the flagged content across categories such as misinformation, adult content, and PII exposure. It generates confidence scores and a PASS, REVIEW, or REMOVE recommendation, helping moderation teams prioritize and escalate unsafe or non-compliant content efficiently.

You can configure thresholds for content moderation recommendations to determine whether flagged content receives a PASS, REVIEW, or REMOVE status. By adjusting these sensitivity settings, the system applies your specific policy boundaries to the per-category confidence scores generated during the scan of user-generated text or mixed media.