Security & Compliance
Your data, our posture
PromptShield processes the inputs you submit for analysis. This page states, plainly, what we handle and what we do not claim.
What we handle
AI app descriptions and untrusted input samples used to produce an injection-risk report.
Data handling commitments
- Submissions run the product pipeline and are retained only as long as needed for your audit log (paid tiers) or until you delete the run.
- We apply access controls consistent with GDPR Art. 32 (security of processing) where personal data is processed.
- BYOK keys (Enterprise), when offered, are stored server-side only and never exposed to the browser.
Honesty rule: PromptShield flags likely injection and jailbreak patterns. We do not guarantee a safe model, full coverage, or that production abuse will never succeed. We do not claim guarantee / 100% / never miss.
Our compliance posture
- PromptShield is decision-support, not a law firm, clinic, or certified auditor.
- For binding advice, consult a qualified professional in the relevant domain.
Subprocessors & payments
- Payments are processed by Waffo Pancake (merchant of record).
- See Privacy and Terms for full terms.
refs: OWASP LLM prompt-injection cheat sheet (LLM01) · hardening checklist