1. Build the Catalog
The catalog lists each real content family as its own Dataset. Articles and products stay separate from directory records, FAQs, and other reusable collections.
The Three-Part Digital Karma Flow
The site names what it publishes, points AI to the right files, and checks the finished result across the full sitemap.
2. Publish the Discovery Layer
The manifest and catalog tell AI where the useful information lives. The health, score, federation, LLM, sitemap, and crawler files keep that discovery system current.
3. Prove It on Every Page
A full-sitemap audit checks that every canonical page carries the same DataCatalog and Dataset identities. A passing receipt records what was actually tested.
The Four Digital Karma Levels
Each level builds on the one before it. The catalog gives every level a concrete job, and public checks show when the site has earned it.
AI-Ready
The site is organized well enough to describe its real content without inventing a generic catch-all Dataset.
How it qualifies
- Canonical pages return clean 200 responses and appear in the current sitemap.
- Organization, WebSite, and page identities use stable Schema.org IDs.
- The catalog can name one Dataset for every populated reusable content family.
Catalog result: the inventory matches what the site actually publishes.
Signal Present
The site publishes a connected discovery layer so AI does not have to guess where identity, content, health, and relationship data live.
How it qualifies
- The manifest points to the current public discovery files.
- The catalog publishes one DataCatalog and the real Dataset inventory.
- The health and score files agree on the same site and version. The federation file, LLM summary, sitemap, and crawler rules do too.
Catalog result: AI can find the Dataset list from the public discovery files.
Verified & Trusted
The site has moved beyond publishing files. It has tested whether the catalog is present in the HTML that AI and search crawlers actually fetch.
How it qualifies
- Every canonical sitemap page renders the catalog's stable DataCatalog and Dataset IDs.
- Each page connects its main subject to the matching Dataset.
- The full-sitemap audit finishes with zero failures and publishes a current rendered-schema receipt.
Catalog result: the Dataset list is verified across the rendered site, not merely stored in one JSON file.
Federation Member
The verified site is connected to its real constellation peers and can add useful network context without giving up its own identity.
How it qualifies
- The federation file names real peers at their declared canonical origins.
- Related-site relationships are current, intentional, and checkable from both sides when the relationship requires it.
- The manifest and health state stay aligned after the site changes. The catalog, rendered receipt, and federation version do too.
Catalog result: the site's verified Dataset identity can be understood beside the related sites that strengthen its subject context.
These are plain-language progress levels. Formal compliance tiers and score-based badges are defined separately in the current standard.
What This Means for a Site Owner
The benefit is not another badge. The benefit is knowing what AI can find and having proof that the setup works.
AI Gets the Site's Real Inventory
The catalog names the content families and resources worth understanding instead of leaving AI to infer the whole site from page copy.
Developers Get a Clear Finish Line
Each level has public checks, so the build can end with evidence instead of ending when the files merely exist.
Owners Can See What Comes Next
The audit shows whether the gap is catalog coverage, rendered proof, or federation relationships. That turns a vague upgrade into a useful next step.
Website Audit
Want to see which level a real site has earned?
Paste any site URL into the audit. It compares the discovery files, catalog, rendered proof, and federation details with the current standard.
When You Need the Full Technical Rules
The standards page defines the exact v8.2 requirements. Start there when you are building, validating, or maintaining the protocol files.