Knowledge Base
Building Federation Endpoints for v8.0
Step-by-step guide to building the required Digital Karma Federation artifacts and root discovery files for a real v8.0 implementation.
Start with the Required Files
| Artifact | Purpose | Owner Benefit |
|---|---|---|
/ai/manifest.json | Primary discovery document | Defines the site clearly for AI systems and developers |
/ai/health.json | Freshness and readiness signal | Makes maintenance status visible |
/ai/catalog.json | Structured resource inventory | Surfaces datasets and machine-readable assets |
/ai/karma.json | Quality and trust posture | Shows how completely the site implements the standard |
/ai/federation.json | Relationship and topology document | Connects the site to peers and supporting properties |
/llm.txt | Plain-text AI summary | Gives language models a fast, readable overview |
For v8.0 you also need /robots.txt and /sitemap.xml at the site root. They are required discovery files even though they are not part of the six-artifact set. Every artifact should also carry protocol_version/lifecycle fields, and none of them should be llm.json -- that endpoint is retired.
Recommended Build Order
- Set the canonical domain and HTTPS rules first. AI discovery is weaker when the site still splits between versions of the domain.
- Create
manifest.json. This becomes the main reference for the rest of the machine-readable layer. - Add
health.jsonandcatalog.json. These tell agents what is current and what resources exist. - Publish
karma.jsonandfederation.json. This is where trust, peer relationships, and standard maturity become explicit. - Add
llm.txt,robots.txt, andsitemap.xml. Keep the last two at the root, never inside/ai/. - Add Schema.org to primary pages. The protocol is stronger when page-level meaning matches the endpoint layer.
Common Mistakes That Weaken Discovery
- Putting
robots.txtorsitemap.xmlinside/ai/. - Using relative or inconsistent URLs in the endpoint files.
- Publishing a version signal that does not match the actual file placement rules.
- Leaving the endpoints technically valid but semantically thin, so the site still says very little about itself.
- Forgetting to update the discovery files after new content ships.
What Owners Get When the Endpoints Are Done
A complete endpoint layer gives owners something far more useful than a compliance badge. It creates a reusable AI discovery surface that can be audited, upgraded, and rolled out across multiple domains. It also makes the site easier for coding agents, research tools, and AI search systems to understand without reverse engineering the entire front end.
For the full plain-language picture, read the v8.0 protocol explainer.