Security overview
How the product handles your data, encryption, access control, audit evidence, retention.
Trust Center
PathReader checks your employees' AI activity on their own devices. Prompts and files go from the device straight to your AI provider, under your account, and responses come straight back. PathReader receives evaluation metadata, not the content: which rules matched, when, and on which machine.
How the product handles your data, encryption, access control, audit evidence, retention.
Current status of each security framework. We publish status, not dates.
Who processes customer personal data, and what never reaches PathReader.
No. They go straight to your AI provider under your account. PathReader receives evaluation metadata, not the content.
No. When you ask PathReader to extract rules, it uses an AI provider key you configure; there is no PathReader-key fallback.
No. Rules are applied by deterministic policy evaluation: your published rules decide whether a rule matched, not a model. Your team approves every proposed rule before it takes effect.
Data is encrypted in transit and at rest, each workspace is isolated, and every sign-in requires multi-factor authentication. The Security overview has the details.
Compliance status lists the current status of each security framework.
We invested first in an architecture that keeps your employees' prompts and files out of our systems entirely; they go directly to your own AI provider. SOC 2 Type I and Type II are next, and we'll publish dates on Compliance status once they're committed.
Sub-processors lists who processes customer personal data, and what never reaches PathReader.
Released documents are available in the gated security package section after you accept a non-disclosure agreement.
Security documents shared under NDA, as they’re released.
Write to security@pathreader.ai.