I built a small set of open-source schema packages after repeatedly seeing production ecommerce sites where JSON-LD technically existed, but product identity and attributes drifted enough to confuse AI systems.

Most “AEO” tooling focuses on monitoring outcomes. This work focuses on the integrity of the underlying product data itself: stable identity, consistency across pages, and explicit handling of conflicting or stale fields.

These packages started as internal tooling while building RankLabs, but I pulled them out into standalone, open-source libraries so developers can inspect and reason about the implementation.

Packages: • @ranklabs/schema (core) • @ranklabs/schema-next • @ranklabs/schema-hydrogen

The goal isn’t SEO automation, but making product data more predictable for LLMs and other machine consumers.

Would love feedback, especially from people dealing with large catalogs or headless commerce setups.