How this actually works
A single-developer product, not a company with a legal department -- this describes real behavior, not a template. See privacy for what's stored, and email directly with any question before relying on anything here.
The service
One incident is fully usable for free -- recording evidence, AI drafting, everything but publishing. Publishing a postmortem, or creating a second incident, requires an active subscription. Content is AI-assisted, never autonomous: a draft is never published without a named human clicking approve, and the database itself enforces that, not just this page's promise.
Your content
You own the incident evidence and postmortems you create. Making a published postmortem public is your own explicit choice, reversible at any time from that incident's settings. Evidence and drafts you keep private are never shown to anyone but you.
Billing and cancellation
Card subscriptions (via Stripe) can be cancelled anytime from the Stripe customer portal, linked from your account settings -- self-serve, no email required. Manual UPI/wire payments are reviewed and approved by the founder directly; to cancel a pending manual claim before it's approved, use the withdraw option on that claim, or email directly. There is no automated refund system for manual payments -- if something goes wrong with one, email directly and it will be handled personally, not by a policy document promising a specific outcome in advance.
No warranty of fitness for any particular purpose
This is an active, single-developer product under continued development, not a finished commercial product with support-level guarantees. The grounding mechanism (every claim cited or marked unsupported) is real and verified -- see how it works -- but a postmortem it drafts is a starting point for your own review, not a substitute for it.
Changes
If these terms change in a way that matters, it'll be reflected here and, for anything affecting an active subscription, communicated by email -- not silently updated with no notice.