Your data stays yours.

We believe Sentience will change how people work with AI. That conviction is exactly why we take the protection of your data more seriously than most. You are handing us something deeply personal: your email, your messages, your meetings, your screen. We have a fundamental responsibility to secure it, and to make sure it stays yours.

Security status: SOC 2 Type 1 complete. Google CASA Tier 2 verified 2026. Visit the Trust Center

Encrypted at every layer.

AES-256 + TLS encryption. Your data is encrypted at rest with AES-256 and in transit with TLS. The keys live in AWS Key Management Service, and the storage that holds your files refuses any upload that is not encrypted with them.

No one sees or can train on your data.

Zero retention. Every AI provider we use is under a zero-data-retention agreement: they may not keep, log or train on your content. We never train a shared model on your data. Only your Sentience is trained from it, and only if you opt in.

You own your digital self.

Ownership. Everything you bring into Sentience stays yours. We claim no ownership over your digital self. Our Terms of Service recognize your data and your personal model as your property, and you can delete them any time.

Independently verified.

SOC 2 Type 1

Complete. An independent auditor examined our security controls in October 2026. Customers can request the report from the Trust Center.

Google CASA Tier 2

Verified 2026. A Google-authorized lab assessed the application against the OWASP ASVS standard, the bar Google sets for access to Gmail and Drive. We are reassessed every year.

You decide what it sees, says and does.

1. You choose what's connected. Gmail, Calendar, iMessage, Slack, meeting audio and screen capture connect only when you turn them on. Disconnect any of them whenever you want, and the data it brought in is cleared.
2. Capture is filtered on your Mac. Screen capture skips any app you exclude, and Sentience's own windows, password managers and private browser windows are blocked before anything leaves your machine. Card numbers and SSNs are scrubbed from captured text. Turn capture off at any time.
3. It asks before it acts. When your Sentience drafts an email, a message or an invite, it asks for your approval before anything reaches another person.
4. It speaks for you only where you say so. Letting your Sentience answer for you, in Slack or on a link you share, is a feature you turn on, and you can turn it off at any time. People you share it with see only the answers it gives them.
5. Delete anything, anytime. Delete a memory, a thread, a connected source, or your whole Sentience from Settings. It leaves our active systems right away, and residual copies in encrypted backups roll off within 30 days.

How it is protected.

Your data is used only for you.

We do not use your data to train models shared across customers or offered to anyone else. If you opt in, we train a personal model for you alone, and we delete it when you turn it off or leave. This is in our Terms, not just a policy.

Every AI provider we use is under a zero-data-retention agreement. They may not keep, log or train on your content. Your content is processed by automated systems. People at Sentience do not read it to operate the service, and production access is restricted and logged. If you send us something in a support request, the person helping you sees it. Beyond that, the only exception is valid legal process, like a warrant or court order.

We never sell your data, and we never share it for advertising.

Where it lives.

Your data is stored with Amazon Web Services in the United States. We are a United States company and process data there. Your files, audio and screenshots live under a storage space that belongs to your account alone, and every request for them is checked against your account.

Stripe handles payments; we see your plan and payment status, never your full card number.

Security is ongoing work.

• Continuous control monitoring. Our cloud and code configuration is checked around the clock, and a control that drifts is flagged the moment it does.
• Threat detection. Account activity, network flows and DNS are analyzed for signs of compromise. High-severity findings reach the team immediately.
• Audit and incident traceability. Every change to our infrastructure is recorded, and the app keeps its own log of sensitive actions.
• Dependency and penetration testing. Every published vulnerability in our dependencies gets a fix opened automatically, and penetration testing is performed annually.
• Vulnerability disclosure. Found something? Email [email protected]. A person reads every report, and confirmed issues are fixed before they are disclosed.

You own your Sentience.

Everything you bring into Sentience stays yours. We claim no ownership over your digital self. Our Terms of Service recognize your data and your personal model as your property. An acquisition does not change that: any successor must honor our Terms or give you the chance to export and delete your data. And when you delete your account, your memories, your files and your personal model go with it.

Read Personal Model Sovereignty

Questions? Ask Sam's Sentience

We're creating a world where everyone can create and own a digital version of their mind.

Read the full Privacy Policy and Terms of Service, or visit the Trust Center. Questions? Email a human.