Privacy policy

What this app does

Heed (Statement Analyser) analyses text you provide using a rule-based engine that runs entirely on your device. The engine examines linguistic patterns associated with research frameworks like SCAN, SVA and Reality Monitoring, and reports a credibility score and per-rule findings.

What data we collect — free tier (current build)

None. The app has no server-side component, does not include analytics SDKs that transmit your input, and does not send any data off your device. Network access is not required for analysis.

What data we collect — future paid AI tier

When AI-powered analysis ships in a future version, paid users will have their statement text sent to Anthropic for analysis through our backend proxy. That data will be processed under Anthropic's data-handling policy. We will not retain a copy on our backend beyond the request itself.

What's stored on this device

We persist a few small things in this app's private storage:

• Your acknowledgement of the first-launch disclaimer (a single boolean).
• The current day's analysis count, used to enforce the free-tier daily cap.
• A list of up to 20 recent analyses — timestamp, your input text, score, and verdict — so you can re-open them from the Recents screen.
• Your theme preference (System / Light / Dark) and crash-reporting opt-in.

All of this lives in the app's sandboxed storage and is not readable by other apps. Use Settings → Delete all local data to wipe everything in one action. See also our data-deletion instructions.

What we don't collect

• We do not collect your name, email, or any personal identifier.
• We do not use analytics SDKs that ship your input text.
• We do not share data with third parties.
• Crash reporting is opt-in only and disabled by default.
• The app's Auto Backup and device-transfer rules exclude everything — Heed data does not sync to your Google Drive or migrate to a new phone.

Reminder about what this app is

Statement Analyser surfaces linguistic indicators consistent with research frameworks. It does not detect lies. It is not evidence and is not a substitute for professional assessment. The empirical validity of the underlying frameworks (SCAN especially) is contested in academic literature — see the in-app glossary under Help for more.

Changes to this policy

If this policy changes in a way that materially affects what data we handle, the disclaimer will re-trigger on next launch and the date above will be updated.

Contact

Email: [email protected]. See also the support page.