Prompt Scan
Prompt Scan
Scan a prompt before it reaches your model.
Static detection for prompt-injection and jailbreak patterns — instruction overrides, system-prompt exfiltration, control-token injection, obfuscation, and tool abuse. Nothing is sent to a model; the scan is local to this request.
0 / 100,000 chars
Static pattern detection only (no behavioral model probing yet — that's on the roadmap). We record only anonymous counts (never your prompt) unless you've opted out in privacy settings.