Your original data stays on your device.

Layrin’s Pro workflow follows a clear boundary: originals and token mappings stay on the local device; after review, protected content may transit through Priova Intelligence’s managed AI service to the AI provider; restoration happens locally.

Data flow text equivalent: local protection; user review; Priova managed service; AI provider; tokenized response; local restoration.

Layrin performs the sensitive-value detection, tokenization, review, correction, and restoration workflow locally in the desktop application.

For Layrin Pro, Priova Intelligence’s authenticated managed AI service temporarily processes the reviewed protected context, protected instruction, and tokenized AI response so the provider can generate a result and return it to the desktop app. Original text, local token mappings, and restored content stay on your device. Priova does not log or persistently retain managed-AI content payloads. See the Privacy Policy for provider and retention detail.

Layrin shows the protected version before sending. You can inspect tokens, keep a value original where appropriate, manually hide additional text, and correct classifications where supported. Sending is a deliberate user action. Automated detection can miss information or classify it incorrectly. The Review step lets you inspect the protected version and manually hide anything else you consider confidential.

Tokens replace original values in the protected text. The mapping between tokens and original values remains local. OpenAI sees tokens, not the mapped originals. The returned response contains tokens, and Layrin uses the local mapping to restore the response on your device. Repeated values may use consistent tokens where supported.

Layrin account and managed-access controls use a small set of verified safeguards.

Service providers support hosting, authentication, billing, managed AI routing, security, transactional email, and website measurement where applicable. The complete provider table, including what each provider receives, is maintained in the Privacy Policy.

Website analytics and advertising technologies are limited to the Layrin marketing website and conversion pages. They are not embedded in the Layrin desktop application and are not used to inspect or transmit user document content. Google Analytics, Meta Pixel, and TikTok Pixel require consent before non-essential activation; advanced matching is disabled at launch, and user-entered text fields are not sent as analytics parameters.

Enterprise and customer-managed AI credentials or endpoints are Coming soon / deferred until after the beta. They are not currently available as self-serve Protect or Pro features.

No software is completely secure. Automated detection may miss information, and users must review before sending. AI providers have their own security and privacy terms. Endpoint compromise, device compromise, malware, clipboard tools, screenshots, and user actions can affect confidentiality. Keep Layrin and your operating system updated, and avoid using Layrin on compromised or shared devices where that would be inappropriate.

Use Layrin only with content you are authorized to process. Follow employer and professional policies, review protected text, verify AI output, use qualified professional judgment where required, and do not treat Layrin as legal, medical, financial, or regulatory advice. Report unexpected disclosure or suspicious behavior promptly.

Priova Intelligence investigates reported security concerns. Access may be suspended or tokens revoked where necessary. Affected users may be notified where legally required. Logs may be retained longer when necessary for an investigation, and Priova may work with service providers during an incident.

If you believe you have found a security vulnerability or unexpected data exposure, contact security@layrin.com with enough detail for us to investigate. Please do not access, modify, or retain other users’ information.