Guardme Privacy Policy

Effective date: September 9, 2026

1. Who we are

Guardme is a personal safety app operated by SAFEIM.AI LTD, a company registered in England and Wales (number 17342838), registered office 2nd Floor College House, 17 King Edwards Road, Ruislip, London HA4 7AE, United Kingdom. We are the data controller for the personal data described here.

Where Guardme is offered. Guardme is currently available in the United Kingdom only. If you are reading this from anywhere else, the app is not being offered to you and we are not processing your data as a user of it. If we make Guardme available in other countries, this policy will be updated first - including, where the law requires it, with the name and contact details of a representative in that region.

2. What we collect

Account information: your email address and, if you sign in with Google or Apple, an identifier and (optionally) your name from that provider.
Profile information you enter: first name, last name, phone number.
Emergency contacts: the name, phone number, and/or email address of the people you choose to add. Section 5 explains what this means if someone has added you.
Your phone's address book, if you use it to pick a contact: the app can read your contacts on your device so it can show you a list to choose from. Only the contacts you actually select are sent to us - the rest of your address book never leaves your phone, and we never upload it. You can also type a contact in by hand and not grant this access at all.
Location: your GPS position, captured when an alert is triggered by either button on the home screen, and sent to the contacts notified for that alert.
Live location for one hour after an alert: once an alert has gone out, your phone keeps reporting its position for up to one hour, so the people you asked for help can still find you if you move. This only ever happens after you send an alert. It never runs at any other time, and there is no tracking when you are not in an alert.
- Who can see it. The alert message contains a link to a page showing your latest position. Anyone holding that link can open it, which is the point: the people you are relying on may be reading it on a phone that has never used Guardme. Treat the link as you would a key.
- The link stops working. It expires one hour after the alert, and it stops immediately if you end the sharing yourself. After that the page shows nothing new.
- You can stop it at any time. While it is running, the app's home screen shows a band with the time left and a button that ends it, and on Android there is an ongoing notification you cannot miss. Ending it stops your phone reporting and switches the page off for everyone who has the link.
- What we keep. The positions reported during that hour are held with the incident and deleted with it. We do not keep a location history outside these windows, and we do not track you between alerts.
Incident recordings: audio and video captured during an alert. These are encrypted on your device before upload; only your device holds the key, so neither we nor anyone else can play them back. Your emergency contacts never receive the recording - an alert sends them your message and your location.
Distress detection signals: when you tap Violence detection, our models run on your device and produce a single decision: alert, or no alert. The camera and microphone feed is not uploaded and does not leave your phone.
Subscription status: whether your subscription is active. Payments are handled entirely by the App Store or Google Play; we never see or store card details.

3. Why we process it, and our legal basis

The article numbers below are those of the UK GDPR - the General Data Protection Regulation as it forms part of the law of the United Kingdom, defined in section 3(10) of the Data Protection Act 2018. They are the same numbers as in the EU Regulation, so if you have read a European privacy notice these will look familiar; the law being applied here is the United Kingdom one, and the regulator is the ICO.

Data	Purpose	Legal basis
Email, provider identifier	Creating and securing your account	Performance of a contract, Art. 6(1)(b)
Your name and phone number	Telling your emergency contacts who the alert is from	Performance of a contract, Art. 6(1)(b)
Your emergency contacts' details	Reaching them when you trigger an alert	Contract with you, Art. 6(1)(b); as regards the contact, protection of vital interests, Art. 6(1)(d), and our legitimate interest in operating a safety service, Art. 6(1)(f)
Location at the moment of an alert	Telling your contacts where you are	Protection of vital interests, Art. 6(1)(d); performance of a contract, Art. 6(1)(b)
Incident recording (encrypted, unreadable to us)	Letting you review what happened	Performance of a contract, Art. 6(1)(b)
Distress detection, after you tap Violence detection	Raising an alert when you may be unable to	Your explicit consent, Art. 6(1)(a) and Art. 9(2)(a); see section 4
Subscription status	Managing access to paid features	Performance of a contract, Art. 6(1)(b)
Account emails	Telling you about your own account	Performance of a contract, Art. 6(1)(b)

4. Automatic detection and health-related data

Automatic detection tries to recognise a violent or medical emergency. A conclusion about whether you are in medical distress is data concerning health, which the law treats as a special category and protects more strictly. We handle it on that footing.

Nothing runs until you tap Violence detection. That tap is where you give explicit consent to this specific processing (Art. 9(2)(a)), and you give it separately each time. There is no background listening and no always-on mode. If you would rather not be assessed at all, use Medical emergency instead: it alerts your contacts immediately and runs no detection.
The analysis happens on your device. What leaves your phone is an alert, not a video feed and not a diagnosis.
Where an alert is raised at a moment you are physically incapable of consenting, we also rely on the protection of vital interests (Art. 9(2)(c)) - which is what a personal safety app exists for.

Automatic detection is a machine judgement and it will sometimes be wrong in both directions. It is an addition to your own judgement, never a replacement for it, and it is not a medical device or a medical diagnosis.

5. If someone has added you as their emergency contact

Your details reached us from the person who added you, not from you, and you may never have used Guardme. This section is for you.

What we hold: your name, and a phone number and/or email address. Nothing else. We do not build a profile of you, and you are not enrolled in anything.
What we do with it: nothing at all until that person triggers an alert. Then we send you their message and their location, by SMS and/or email.
Why we are allowed to: the protection of vital interests (Art. 6(1)(d)) and our legitimate interest in operating a safety service (Art. 6(1)(f)). Someone in danger cannot reasonably be asked to collect your written permission first.
How to be removed: ask the person who added you, or write to info@safeimai.com and we will remove you. You do not need an account to ask, and we will not require you to create one.
Your rights are the same as anyone else's - access, correction, erasure, and objection (Art. 21). If you object, we stop, and we do not ask you to justify it.

6. On-device face detection (Google ML Kit)

Guardme uses Google ML Kit's on-device Face Detection API to locate a face in the live camera preview during active recording, so the app can crop the relevant region before running our own detection models. This runs entirely on your device: no image or video is sent to Google, and ML Kit does not transmit your camera feed or any biometric data off your phone.

7. Who else is involved

Your emergency contacts receive your alert, and only when you trigger one. Beyond that, a small number of providers process data on our behalf, each receiving only what their part requires:

Fly.io (a United States company) - hosting for our servers and our database, in their London region. This is where your account, your emergency contacts, your incident history and your encrypted recordings are stored.
Google Firebase (Google Ireland Limited) - sign-in and account authentication.
Apple - Sign in with Apple.
RevenueCat, with the App Store and Google Play - subscription status and billing.
Twilio - SMS delivery of alerts.
IONOS (IONOS Cloud Ltd, registered in England and Wales) - delivery of alert and account emails.

The app does not work without an internet connection.

We do not sell personal data, we do not use it for advertising, and we do not share it with law enforcement except where the law requires it.

8. Where your data goes

Your data is stored in the United Kingdom. Our hosting provider keeps our servers, our database and our file storage in its London region, and we have not switched on any other region.

Where it is stored is not the whole picture, and we would rather set out the rest than leave you to assume it. Fly.io, RevenueCat and Twilio are United States companies, so the law treats our use of them as a transfer out of the United Kingdom even where the data itself stays in London. Those transfers are covered: Fly.io participates in the UK Extension to the EU-U.S. Data Privacy Framework, and RevenueCat and Twilio operate under the UK International Data Transfer Addendum to the Standard Contractual Clauses. Google, through Google Ireland Limited, processes within the European Economic Area, which the United Kingdom recognises as offering equivalent protection. IONOS is contracted through IONOS Cloud Ltd, a company registered in England and Wales, so nothing leaves the United Kingdom there at all.

One thing does travel, and it is worth being clear about it: if you add an emergency contact who lives abroad, their name and contact details are stored here in the UK and an alert is sent to them wherever they are. That is the point of the feature, and it happens only when you trigger an alert.

Incident recordings are a special case: they are encrypted on your device before they are uploaded anywhere, so wherever they are stored, their contents are not readable by us or by the provider holding them.

9. How long we keep it

Your account and profile: while your account exists. A lapsed subscription pauses paid features; it does not delete anything.
Incident recordings: until you delete them. They are protected from deletion for 5 days after they are created, and deleted automatically about 12 months after creation - we email you a warning first, and delete roughly 30 days later if you have not saved or deleted them yourself.
Live location positions: kept with the incident they belong to and deleted with it. Any that outlive their incident are cleared within 24 hours of the hour ending.
Inactive accounts: if you have not opened the app for about 12 months we email you a warning, and delete the account if you have not returned roughly 30 days later.
Deleting your account from Profile > Danger Zone removes your profile, emergency contacts and incident history.

Uninstalling the app does not delete your account. Use Profile > Danger Zone if you want your data removed.

10. Security

Incident recordings are encrypted on your device with a key only your device holds. Access to everything else is restricted to what the app needs, enforced by server-side rules rather than by the app asking politely. No system is perfectly secure, and we do not claim otherwise.

11. Your rights

You can access, correct, erase, restrict, port, and object to the processing of your personal data. Where processing rests on consent - the distress detection that follows a Violence detection tap - you give that consent per alert and can simply stop using that button; withdrawing it does not affect what was done before.

Most of this is immediate and in your hands: view and edit your profile in the app, and delete your account and all associated data from Profile > Danger Zone. For anything else, write to info@safeimai.com; we answer within one month.

If we get something wrong, you can complain to the privacy regulator where you live: in the United Kingdom, the Information Commissioner's Office (ico.org.uk). We would rather you told us first, but that is your right and not conditional on asking us. If you do come to us, we will look into what happened, tell you what we find and what we are doing about it, and we will not ask you to give up the right to go to your regulator as a condition of us dealing with it.

12. Children

Guardme is not intended for children. The minimum age is 13, which is the age UK law sets for a service like this. If we make Guardme available in countries whose law sets a higher age - several set 16 - the higher age will apply there and this policy will say so before that happens.

We do not knowingly collect data from anyone below that age. If you are a parent or guardian and believe your child has created an account, write to info@safeimai.com and we will delete it.

13. Changes to this policy

We will post any change here with a new effective date. Where a change materially affects you, we will tell you in the app rather than relying on you to re-read this page.

14. Contact

SAFEIM.AI LTD - info@safeimai.com

Our Privacy Officer is accountable for how this policy is applied, and is who to write to about any of it - a question, a request about your data, a complaint, or a concern that something has gone wrong. Use the address above. We publish the role and not an individual's name, so that the contact point stays valid whoever holds it.