Digital Karma Federation
The Plumbing, Not the Pitch.
An open protocol for publishing standardized machine-readable endpoints, so AI agents reading two different sites find the same structure in the same places.
What it actually is
The Digital Karma Federation specifies a consistent shape for the endpoints a site publishes about itself: what it contains, what content types exist, how to reach the important parts, and how healthy the site is.
Any site can implement it independently. There is no registration requirement and no central authority controlling the network.
/ai/manifest.json... what the site is and what it publishes/ai/catalog.json... the datasets the site exposes/ai/health.json... operational status/ai/karma.json... the site's own scoring signals/ai/federation.json... network relationships to related sites/llm.txt... a plain-text orientation for AI systems
What it does not do
This is the part usually oversold, so here it is straight.
Publishing endpoint files does not by itself create rankings, mentions, or AI visibility. It makes a site legible to machines. That is useful plumbing. It is not a marketing channel, and anyone selling an llm.txt file as a growth strategy is guessing.
What the endpoints do is remove ambiguity. An agent that can read a structured description of your site does not have to infer it from markup. Whether that agent then cites you depends on things nobody controls.
See machine-readable endpoint and llm.txt for the definitions, and data intelligence for how crawler activity is actually measured rather than assumed.
This site publishes them too
Every endpoint listed above is live on this domain. That is the point: the standard is only worth anything if the sites describing it also implement it.