The quiet infrastructure
of trust.

Identity

Every Aletheia account is tied to a real person. Before you see a single profile, you complete a one-time identity check: upload a government-issued ID (passport, driver's licence, or national ID), then take a live selfie. Our partner Persona runs a three-way check — document authenticity, biometric match, and liveness detection.

The entire flow takes about 60 seconds. Once it passes, Persona sends us a single verified flag and a one-way hash of your facial geometry. The ID image itself is never stored in our infrastructure. We can confirm you passed without being able to read who you are.

What we keep:

• A verified-account flag (boolean)
• A tokenised reference ID from Persona (opaque to us)
• The hash of your facial geometry (cannot be reversed to a photo)

What we never keep:

• A copy of your government ID
• A plaintext photo of your face
• Your full legal name or date of birth (unless you voluntarily add them to your profile)

Re-verification is triggered automatically if our fraud model detects an anomaly — for example, if your uploaded profile photos change dramatically over time. In that case, you'll be asked to re-verify in under a minute.

Vouching

The "4 mutual friends" chip you see on a profile isn't decorative. It means four people in your existing network have independently connected with that person on Aletheia — and none of them revoked the connection.

The vouching system works like this: when you add your phone contacts or connect an existing social account, we find the overlap with other verified users. We don't share your contact list with anyone else, and we never message your contacts on your behalf.

You can also actively vouch for someone. If a friend asks you to introduce them to someone you know on Aletheia, you can tap "Vouch for [name]" on their profile. This creates a one-way signal: you're saying this person is who they say they are, and they treated you with respect. You can revoke a vouch at any time, and revoked vouches are hidden but logged for our safety team.

We surface vouch count prominently because it's our best proxy for social proof in a world where AI-generated profiles are increasingly convincing. A profile with twelve mutual connections is simply harder to fake than one with none.

Reporting

One-tap reporting is available on every profile and in every conversation. No friction, no "are you sure?" confirmation dialogs. When you report, the report goes directly to a human reviewer — not a classifier, not an automated queue.

Our SLA for report review is four hours. In practice, during business hours, it is under ninety minutes. We publish a quarterly transparency report showing report volumes, categories, and outcomes. The first edition will be published when we exit TestFlight.

Report categories:

• Impersonation or fake identity
• Inappropriate or abusive messages
• Harassment or threatening behaviour
• Underage user
• Solicitation or spam
• Other

Accounts reported across multiple independent users in a 30-day window are automatically suspended pending review, regardless of the outcome of earlier reports. A single egregious report (threatening behaviour, solicitation of a minor) triggers immediate suspension.

Banned accounts cannot re-register. Because every account is ID-verified, "ban evasion" requires a different government document — a much higher bar than simply creating a new email address.