Finance Secretary — Security & Data Protection

Last reviewed: May 2026
Published by AI Secretary Limited (trading as AISec Technology), company number 17245992, registered in England and Wales.
Security contact: [email protected]

In short. Your financial data lives on your own device, not on our servers. Document classification runs on private AI hardware we own in the United Kingdom — never a third-party cloud. There is no advertising, no analytics, and no tracking. This page explains how we protect your information and how to report a security concern.

1. Where your data lives

On your device. Your transactions, balances, assets, liabilities and uploaded documents are stored in a local database inside the app's sandboxed storage, encrypted at rest by iOS Data Protection. They do not leave your device except for the limited AI-classification step in section 3.
Your API key is held in the iOS Keychain, which is hardware-encrypted on every modern iPhone.
On our server we hold only your account email and an API key — encrypted at rest — on infrastructure we own in the United Kingdom. Nothing else about your finances is stored server-side.

2. How information is protected

Encryption in transit. All traffic between the app and our service uses HTTPS (TLS 1.3), with certificate-chain validation. The app enforces HTTPS-only connections (App Transport Security with arbitrary loads disabled).
Passwords. Stored only as a bcrypt hash — we never see or keep your plaintext password.
Authentication. Sessions use signed tokens that expire and require re-login; mobile sessions additionally carry a hard expiry, so a lost device loses access within a fixed window even without a remote revoke.
Device unlock. You can require Face ID / Touch ID to open the app.
Server-side key storage. The account API key we hold is encrypted at rest using authenticated symmetric encryption.
Abuse protection. Authentication and sign-up endpoints are rate-limited to blunt automated guessing and spam.

3. The private AI classification step

When you upload a PDF or image, the text is read on your device using Apple's Vision framework. Only that extracted text — never the original file, your name, email, or password — is sent over HTTPS to our private AI service at api.aisectech.uk. The service runs on hardware we own, in our own location; it is not a third-party cloud. The text is held only briefly to complete the job, then automatically purged. It is not logged, not used to train any model, and never shared outside our own infrastructure.

4. What we deliberately don't do

No analytics, crash-tracking SDKs, or telemetry in App Store builds.
No advertising and no ad identifiers.
No selling or sharing of your data with third parties.
No using your data to train AI models.
No cross-app or cross-site tracking, and no cookies.

5. What we design against

Our security work is built around a written threat model. The main risks we design for, and the protections that address each:

A lost or stolen device → app sandboxing, encrypted local storage, Keychain-held keys, and optional Face ID / Touch ID.
Interception of network traffic → TLS-encrypted transport with certificate-chain validation.
Exposure of a single account's API key → a key grants access only to that one account's AI usage; no other user's data is reachable with it.
A malicious uploaded document → every AI suggestion is shown to you for review before anything is saved, and the AI cannot take actions on its own.

Like any consumer application, Finance Secretary is not designed to withstand state-level adversaries or an attacker who has already compromised your device's operating system.

6. Ongoing review

We carry out a security review before each major release and re-audit after significant changes — covering authentication, data storage, transport, the AI pipeline, and dependency advisories. Where a review identifies something to harden, we fix it before shipping rather than publish a live list of open issues. We're happy to discuss our practices with security researchers.

7. Reporting a security issue

If you believe you've found a vulnerability or a privacy problem, please email [email protected] with enough detail for us to reproduce it. We'll acknowledge your report, investigate, and keep you updated. Please give us a reasonable opportunity to resolve the issue before disclosing it publicly. We don't currently run a paid bug-bounty, but we're grateful for responsible disclosure and will credit you if you'd like.