product log · NO QUIET CORPSES

Changelog

What changed. What got stronger. What keeps watching.

Zombify tracks AI model lifecycle risk. This page tracks Zombify itself. Reliability upgrades, lifecycle improvements, dashboard changes and product hardening live here.

4 May 2026 · v1.2.0

Lifecycle, authentication and delivery hardening

Reliability upgrade

A major reliability update across Zombify’s model monitoring, dashboard access and email delivery systems. This release strengthens the parts of the product that matter most: trusted lifecycle data, reliable alerts and secure account access.

Access

  • Added secure magic-link dashboard authentication for paid users.
  • Dashboard access is now tied to control of the paid checkout email inbox.
  • Improved session handling across dashboard and watchlist actions.
  • Added stronger server-side checks for account access and watchlist activity.

Email delivery

  • Improved transactional email delivery using a verified domain sender.
  • Moved login emails to a dedicated Zombify sender address.
  • Updated authentication email copy so dashboard login feels branded and clear.
  • Improved sender authentication so login links arrive more reliably.

Lifecycle accuracy

  • Strengthened lifecycle verification rules for monitored AI models.
  • Prioritised official provider sources when lifecycle signals conflict.
  • Improved handling of third-party availability signals.
  • Expanded lifecycle visibility across active, at-risk, deprecated, zombie and retired models.

Recommendations

  • Improved replacement recommendation safeguards.
  • Replacement suggestions now favour models that are currently suitable for migration.
  • Refined migration paths for legacy OpenAI and Anthropic models.
  • Added stronger checks so replacement guidance stays aligned with live model status.

Watchlists

  • Tightened watchlist rules for monitored model stacks.
  • Dashboard watchlists now focus on models that are still suitable for active monitoring.
  • Improved model validation before a model can be added to a watchlist.
  • Kept the public model checker broad so users can still search living and dead models.

Transparency

  • Added a public changelog for clearer product updates.
  • Improved release visibility for reliability, lifecycle and product changes.
  • Added this page so Zombify’s own evolution is easier to follow.

28 April 2026 · v1.1.0

Registry review and live model source of truth

Registry upgrade

Introduced a stronger model registry workflow so Zombify can review, classify and update model lifecycle records as providers change their model line-ups.

Registry

  • Added live registry review tools for model candidates.
  • Added provider, status and model search filters for registry management.
  • Added review actions for model lifecycle decisions.
  • Improved source notes and proof links for reviewed lifecycle records.

Ingestion

  • Improved provider-model registry handling across OpenAI, Anthropic, Google and OpenRouter.
  • Added clearer separation between discovery signals and final lifecycle judgement.
  • Improved official-source notes for model candidates.
  • Strengthened the foundation for ongoing lifecycle monitoring.

24 April 2026 · v1.0.0

Zombify launch build

Initial launch

The first public version of Zombify: a model lifecycle checker, live Graveyard, paid watchlist and shareable status page for developers who do not want to learn about model deprecations from their users.

Core product

  • Launched the public model checker.
  • Launched the model Graveyard.
  • Added paid watchlists for up to 10 models.
  • Added shareable status pages for monitored model stacks.
  • Added Stripe checkout for lifetime access.

Brand

  • Launched the Zombify visual system, homepage, dashboard and mobile navigation.
  • Added the undead model monitoring positioning across the homepage and product UI.
  • Introduced the first version of Zombify’s lifecycle monitoring experience.

Watch the models before they watch you suffer.

Check your model stack, scan the Graveyard, then add the living ones to your watchlist. Do not find out from a user.

Check your stack →