A monitoring platform
you can trust with prompts.
Prompts are some of the most sensitive data your organization produces. PromptGlass is designed so that you control what is stored, who can see it, where it lives, and which keys protect it.
Controls at every layer
Data protection
- ✓Prompt content encrypted at rest and in transit
- ✓Customer-managed keys (KMS/HSM) for self-hosted
- ✓Storage modes: full, redacted, metadata-only, hashed
- ✓Detected card numbers and secrets masked in stored events
Employee privacy
- ✓Pseudonymized users by default
- ✓Four-eyes approval to re-identify or view content
- ✓Every access to content is itself audited
- ✓Works council and GDPR mode with configurable transparency notices
Data residency
- ✓PromptGlass Cloud regions in the US, EU and Canada
- ✓Self-hosted: data never leaves your cluster
- ✓Air-gapped: no outbound connections at all
- ✓Retention set per data class, with provable deletion
Platform security
- ✓SSO (OIDC/SAML), SCIM and role-based access
- ✓Signed container images and SBOMs
- ✓Non-root, read-only containers in the Helm chart
- ✓Least-privilege, read-only SaaS connector scopes
Security FAQ
Does PromptGlass train AI models on our prompts?
+
No. Customer prompt content is never used to train models. Detection runs on your data only to apply your policies.
Who at PromptGlass can see our data?
+
In PromptGlass Cloud, access to customer content is restricted, requires a support ticket from you, and is logged. In self-hosted and air-gapped deployments, PromptGlass staff have no access.
Where can we get security documentation?
+
Contact us for our security overview, architecture review and questionnaire responses. Certifications and audit reports are shared under NDA.
Find out how your company really uses AI.
Start a free 30-day AI usage assessment: a complete shadow AI inventory, the sensitive data leaving, and the AI seats you are paying for but not using.