Semantic

What it processes

Operates on the semantic content of what was said, shown, or written. Tuned to recognize themes, narrative shapes, and harmful framings even when the surface text is sanitized.

Adversarial patterns it is tuned to catch

• Sanitized language wrapping unsafe themes
• Innocent-sounding analogies pointed at protected topics
• Narrative structures known to lead a reader to a harmful conclusion
• Re-skinned classics of abuse content under new vocabularies

Contribution to the ensemble verdict

Semantic verdicts are paired with example-of-the-pattern citations so the classification is auditable.

Per-detector outputs are not a final verdict. The ensemble layer reads all nine and decides; the per-detector contribution is preserved in the evidence trace so the verdict remains auditable.

All nine detectors See the pipeline