Built for Indian FamiliesHonest About What Is Built

Planned for launch. Parents will create a profile for each child, set the rules that apply to them and attach an emergency contact. Watching a child's app usage or blocking content on their handset needs Android controls the app does not have, so those are not part of the first release.

• Per-child profiles
• Rules set by the parent
• Emergency contact per child
• App-usage limits: not in the first release

One action raises an alert and keeps adding to a location trail until you mark yourself safe. The alert endpoints are built; the app screens that drive them are not finished, and nothing yet passes the alert on to anyone else. There is no automatic audio recording — recording another person is a legal question, not a feature toggle, and we will not ship one without a consent flow around it.

• One action raises an alert
• Location trail until you resolve it
• History of past alerts
• Alerting your contacts: not built yet

Checks written around the scams that actually target India — fake UPI collect requests, KYC lures, urgency and fear wording, and links that look risky. The rules run on our servers and are hand-written patterns, not a machine-learning model. We would rather say that than call them AI.

• UPI fraud patterns
• Phishing link checks
• Risk score with the reasons listed
• Hand-written rules, not AI

A posture check of your Android device: root, permissions that are riskier than they look, and apps holding accessibility or screen-overlay access. The server side that records a report and keeps its history is built; the app does not run any of these checks on the device yet.

• Root and risky-permission review
• Accessibility & overlay audit
• Report history kept per device
• On-device checks not built yet

Once a capture is saved, the server will not delete it — that refusal is written into the code, not into a policy document. The server works out each file's SHA-256 fingerprint from the bytes it receives rather than trusting the app, and a later check re-reads the stored file and works it out again, so a file altered after the fact no longer matches and is reported as altered. Each capture is stored with its time, a short reference ID, optional GPS coordinates and an audit-trail entry, and every evidence route sits behind a consent check. Files are encrypted by us before they are stored — each with its own key — so the copy held by our storage provider is not readable as your file, and a stored file that has been altered fails to decrypt rather than being handed back. The app screens for capturing and browsing evidence are not finished. We still do not call the vault court-ready: that is a judgement for a court, not a feature we can ship.

• Saved captures cannot be deleted
• Time, reference ID, optional GPS
• Audit trail on every capture
• Consent required before access

A single action raises an alert and keeps updating your location until you resolve it. Trigger, location-update, resolve and history endpoints are built; the screen that calls them is not, and nothing yet forwards the alert to anyone else. That is also why no activation time appears here — nothing has been measured yet.

• One action to raise an alert
• Location updates until resolved
• Past alerts kept in history
• Forwarding to your contacts: not built yet

One thing first: the app does not read your inbox. Android's SMS permissions were deliberately removed and a test keeps them out, so you share or paste a message in when you want it checked. The check is a pattern rule set on our servers — UPI wording, phishing links, urgency tactics — returning a risk level and the reasons behind it. It needs a connection; there is no offline mode.

• No inbox access — you choose what to check
• UPI & phishing patterns
• Risk level with reasons
• Needs a connection

Planned for launch: a lock on the Evidence Vault and other sensitive screens, using the device's own fingerprint, face unlock or PIN. No screen in the app is behind it yet, so nothing here treats it as protection today.

• Fingerprint or device unlock
• PIN fallback
• Locks the vault and sensitive screens
• No screen uses it yet — wired up before launch

Six language files ship with the app — Hindi, Tamil, Telugu, Bengali, Marathi and English. Nothing in the app reads them yet: every screen still shows text written straight into the screen. Connecting the app to these files is launch work, so the app is not usable in these languages today.

• 6 language files shipped
• Hindi, Tamil, Telugu, Bengali, Marathi, English
• No screen reads them yet
• Not usable in these languages today

Planned for launch. Report a scam number, message or site in a few taps, and see what other members have reported. The feed is empty and stays quiet until there are members filing reports — so we are not claiming intelligence we do not have yet.

• Report a scam in a few taps
• See what other members report
• The feed is empty today
• Grows only as members file reports

Planned for launch. A critical alert should interrupt you and everything else should not. The priority tiers are designed; the notification channels behind them are not implemented yet.

• Critical alerts break through
• Routine updates stay quiet
• Channels not implemented yet