Privacy

The product

IncidentScribe is a Mac-native postmortem drafter that runs entirely on-device using Apple’s Foundation Models framework. Incident drafts — the artefacts you import, the timelines extracted from them, the postmortem sections you produce — never leave your Mac.

The commitments:

This website

This marketing site (incidentscribe.app) is hosted on Cloudflare Pages and collects two things:

  1. Your email address, only if you submit the signup form. Stored in a Cloudflare Workers KV namespace alongside the timestamp of signup. We email you exactly once — when IncidentScribe v1.0 ships on the Mac App Store. To delete your record, email [email protected] with “unsubscribe” in the subject.
  2. Anonymised analytics via Google Analytics 4 (gtag.js). IP-address anonymisation is enabled (the last octet is dropped at Google’s edge before logging). We see page views, traffic sources, browser/OS aggregates, and our two conversion events: sign_up (newsletter signup) and app_store_click (App Store CTA tapped). We do not enable Google Signals, ad personalisation, or cross-property tracking.

What we do NOT do:

To opt out of GA tracking: install Google’s opt-out browser add-on, or any standard tracker-blocking extension (uBlock Origin, Privacy Badger). The site works fully without GA.

The app vs. the website

Critically: the analytics on this website do not extend to the IncidentScribe macOS app. The app ships with zero analytics SDKs, zero telemetry, zero remote logging. Its privacy posture is enforced by the App Sandbox and is independent of anything on this domain. The website’s GA is for marketing-funnel measurement only.

Server logs

Cloudflare records standard HTTP request logs (timestamp, IP, URL, response code) for traffic served from this domain — the same as any web hosting provider. We do not export these logs, do not feed them to a third party, and do not retain them ourselves beyond Cloudflare’s default retention.

Contact

[email protected] — for questions, deletion requests, or anything else.