To view monitoring data, tap the child menu (โฎ) on the Dashboard and select "Activity Reports".
Every monitoring feature is independently toggleable and disabled by default. You decide what's appropriate for your family. Temporary data like SMS and notification content is automatically deleted after 7 days.
When enabled, you can see who your child has called and who has called them โ including the phone number, contact name (if saved), whether it was incoming or outgoing, duration, and timestamp.
When enabled, you can view text messages sent and received on your child's device โ including the phone number, contact name, direction (sent/received), and the full message text. SMS data is kept for 7 days then automatically deleted.
When enabled, you can see notifications your child receives from messaging apps (such as WhatsApp, Instagram, etc.) โ including the app name, contact name, and a message preview. Notification data is kept for 7 days then automatically deleted.
Keyword monitoring is currently an experimental feature that is being actively developed and refined. Detection accuracy may vary, and you may see occasional false positives. We recommend using it as one tool among many for keeping your child safe, not as a sole safeguard.
ChildSafe includes an on-device keyword monitoring system that scans text content for concerning language. The curated keyword database covers 12 categories with thousands of keywords and slang variants:
- Self-harm & Suicide โ phrases indicating your child may be in distress
- Grooming โ patterns used by online predators
- Bullying โ harassment and aggressive or threatening language
- Sexual Content โ inappropriate or adult material
- Drugs & Alcohol โ references to substance use
- Violence โ fighting, weapons, and threats
- Profanity โ swear words and crude language
- Radicalisation โ extremist language and recruitment patterns
- Financial Exploitation โ scams, gambling, and predatory financial content
- Gaming Predatory Practices โ in-game grooming, trading scams, and exploitation
- Body Image โ disordered eating, unhealthy body image content
- Misogyny โ gender-based harassment and incel-related language
Each category can be individually toggled on or off in the Keyword Settings screen. Keywords are assigned a severity level, and categories like Self-harm and Grooming are marked as critical by default. When you enable critical keyword alerts, detections of the most serious keywords will trigger an instant push notification on your phone. The notification tells you that concerning content has been detected and which app it was found in. You can then review the full details in the parent app.
Lower-severity detections are logged and visible in Activity Reports but do not trigger a push notification. You can configure which categories are treated as critical in the keyword monitoring settings.
Keyword scanning runs on the device. When a keyword matches, the matched word, the app it appeared in, and a short snippet of surrounding text (about 300 characters) are reported back to ChildSafe so you can see what was flagged. The snippet is also forwarded to an AI service (Anthropic) that decides whether the match is a real concern or an everyday usage like a song lyric or a school project โ this is what stops you being bombarded with false alarms. No names, accounts, or device IDs are sent. The full message is never stored. See the Privacy page for full detail.
In addition to the curated keyword database, you can add your own custom words or phrases to monitor. This is useful for tracking specific terms that matter to your family โ for example, a particular nickname, a local slang term, or a word related to something your child is dealing with.
To manage custom keywords, go to the child menu (โฎ) on the Dashboard, select "Activity Reports", then tap "Keyword Settings". Scroll down to the "Custom Keywords" section where you can:
- Add a word or phrase โ tap "Add Word", enter the word (2โ50 characters), and optionally assign it to a category
- Delete a custom keyword โ tap the delete icon next to any custom word to remove it
Custom keywords are synced to your child's device and scanned alongside the curated database. They appear in Activity Reports with the same detail as built-in keyword matches.
When web filtering is enabled, ChildSafe logs which website domains your child visits (e.g. "youtube.com", "roblox.com"). This shows domain-level activity only โ not full page URLs or page content. Browsing logs are retained for up to 90 days.
When enabled, ChildSafe scans images that arrive on your child's device โ camera photos, images saved from messaging apps like WhatsApp or Telegram, screenshots, and downloads โ for nudity. If something is detected, the image is automatically blurred on the device, and an alert is sent to your phone.
The scan runs entirely on the device. No image ever leaves the phone โ not to ChildSafe, not to anyone. The detection model lives locally on the device and processes each image in milliseconds.
You can enable image safety scanning and adjust its sensitivity from the Keyword Settings screen (child menu (โฎ) โ "Activity Reports" โ "Keyword Settings"). Image detections always trigger an alert โ they don't go through the AI false-positive filter that text keyword alerts use.