Security model
No cloud. No exceptions.
Triage is built around a single constraint: your data never leaves your device. Everything — storage, inference, decision history — runs locally in an encrypted vault only you control.
Core security properties
How we protect your data.
AES-256 on-device vault
Every email, task, preference, and decision Triage touches is encrypted with AES-256-GCM. The key is derived from your device's secure enclave — only you can decrypt it.
Zero network calls
Triage does not phone home. After the initial OAuth handshake with your email provider, all processing is local. We receive no telemetry, crash reports, or usage data.
On-device inference
The AI model that powers Triage runs entirely on your device. No request is sent to an AI API. No prompt data leaves your vault under any circumstances.
Local audit log
Every action taken in autonomous mode is written to a tamper-evident local audit log. You can inspect, export, or delete it at any time from the settings panel.
Offline license
Your Pro or Family license is a cryptographically signed file stored on your device. There is no activation server to be unavailable or to record your usage.
No account, no breach surface
We hold no user accounts, no email addresses, no payment tokens. There is no Triage user database to breach or subpoena. We structurally cannot see you.
Technical specifications
Security at a glance.
| Property | Details | Status |
|---|---|---|
| Vault encryption | AES-256-GCM, key in device secure enclave | Enforced |
| Data-at-rest | All vault content encrypted, never plaintext on disk | Enforced |
| Network calls (processing) | Zero — all inference is on-device | Blocked |
| OAuth scope | Minimal read/send scopes only; tokens stored encrypted | Enforced |
| Telemetry | None — no analytics SDK, no crash reporter | Disabled |
| AI inference location | On-device model only | Local |
| License verification | Local signature check, no activation server | Offline |
| Audit log | Local tamper-evident log, user-exportable | Always on |
Responsible disclosure
Found a vulnerability?
Reporting security issues
If you discover a security vulnerability in Triage, please disclose it responsibly. Do not open a public GitHub issue for security findings.
Email us at triageaiinbox@gmail.com with a description of the issue, reproduction steps, and any relevant technical details. We will respond within 72 hours and work with you to patch and credit the finding.
We do not currently operate a bug bounty program, but we take every report seriously and publicly credit researchers who help us improve Triage's security.