On-device child safety

Protect them now.
Prepare them for later.

AI detects the risk. You decide what happens next. Your child learns from the moment.

Most child-safety products restrict access or report harm after it happens. RoseShield looks at the interaction itself and can block harm before it reaches your child. Unsafe outgoing content can be paused and explained, giving your child a chance to change it before it reaches someone else.

GroomingBullyingSextortionSelf-harm& 34 more

See more demonstration videos →

Private by architecture.

On the device

Messages, images, voice and video are analyzed on the phone, tablet or laptop — not in the cloud.

Not another archive

RoseShield does not create a separate store of a child’s conversations.

AI detects. Humans decide.

AI detects and explains. You, or the person you designate, decide what happens next.

One

Protect
them now.

When something serious is found, it does not reach your child — and it does not sit in a report until the end of the month. It comes to you, with the reasoning attached, while there is still time to act.

01Harm detectedOn the device, as it arrives or as it is sent.
02StoppedIt does not reach your child.
03You are toldWith the reasoning attached, not a bare alert.
04You decideNothing is auto-reported. No one is contacted for you.
Two

Prepare them
for later.

Protection is only half the job. The rest is helping your child make better choices, giving you time to step in when needed, and showing you the longer pattern without exposing their private life.

A message stopped on a child’s phone, with the reason shown and the option to edit it
The pause

A second chance, in the moment

When your child sends something cruel or unsafe, RoseShield stops it before it reaches anyone, explains the concern in a sentence, and gives them the chance to put it differently.

You are told it happened. They get a chance to reconsider, and repeated practice at making the better choice themselves.

A parent reading an alert on her phone while her children play nearby
Your alert

Time to help, instead of time to react

When something genuinely needs an adult, you know while there is still time to do something about it — not in a report at the end of the month.

Problems that once stayed hidden can surface while there is still time to act. You can talk to your child, understand what is happening, and decide what to do next.

RoseShield Signals

The longer view

Individual incidents tell you about a day. Signals shows how your child’s relationship with their phone is changing across weeks and months. What looks healthy. What has changed. What may deserve a closer look. A few lines on a Sunday, in plain language — weekly, monthly, or never.

A RoseShield Signals weekly summary on a tablet

And it changes as they grow up.

You decide when RoseShield loosens — not an algorithm, and not a birthday.

10More protectionLimits stay firm; higher-risk content is blocked outright.
13More judgmentYou loosen lower-risk limits as they are ready.
16More independenceMost of it becomes advice, with hard limits kept for serious risks.
Channel partnerships
Five African countries

South Africa, Kenya, Uganda, Cameroon and Ghana — each led by a local organization.

Pilots signed
Kampala

The first two school pilots, with the Online Safety Foundation Uganda.

Framework agreed
United Arab Emirates

A cooperation framework with the Emirates Safer Internet Society.

Broad pilot
RoseShield Messenger

Entering broad pilot testing with a major international children’s organization.

See partnerships and deployments →

How it works

Five steps, all on the device, as content is sent or received.

01

Understand

Works out who is talking to whom, and whether a child is involved.

02

Analyze

Reads text, images, voice and video together.

03

Detect

Looks for 38 kinds of harm, including patterns that build slowly.

04

Score

Rates how serious it is, and why it was flagged.

05

Protect

Holds, warns or escalates based on the level of risk and your settings.

How do you want to use RoseShield?