Blog: ODIN Catalog

🔁

Agents That Review Their Own Work

A one-shot LLM gives you an answer, not a reason to trust it. ODIN's agentic metadata review puts two models in a loop; a proposer drafts every element's description, classification, vocabulary, and PII flags; a reviewer critiques the draft against the full dataset and pushes back until it runs out of objections. With long-term memory, live SSE streaming, and a human who signs off at the end.

Read the engineering deep dive →

Standards 3 June 2026 · 18 min read

ODRL Under the Hood

A technical deep-dive into the W3C ODRL standard; the information model, the JSON-LD wire format, the constraint algebra, and the four policy types. How six building blocks (Policy, Asset, Party, Rule, Action, Constraint) combine into governance that machines can evaluate, enforce, and audit without human interpretation.

Read the deep dive →

Design 31 May 2026 · 18 min read

The ODIN Design Manifesto

The data governance tools we use today were designed for a world where having metadata was enough. ODIN is designed for a world where understanding metadata is the only thing that matters. Covering meaning vs. naming, standards as architecture, lineage as a live graph, AI contracts, and the eight principles every design decision is tested against.

Read the manifesto →