Text Safety Analyzer (multi-model)
Paste text or a prompt below. The system runs heuristics + multiple models (toxicity/harm + URL detection) and returns
explainable flags. This is a detection tool—human review recommended for any enforcement action.