The knowledge graph is a medium. Every AI answer is a composition drawn from it. SPXI Protocol is the craft of inscribing your entity into that medium durably — so the composition is yours, not the machine's. Every claim anchored at a content-derived identifier — AXN in Alexanarch, with historical Zenodo DOIs preserved in the resolution index.
§ 07Our own position, in the terms of our own instruments
The Institute measures provenance erasure and semantic rent. It also sells entity deployment. Rex Fraction is the commercial voice. Lee Sharks is the archival authority. The instruments are CC BY-SA and usable by people who never hire us, and where a measurement of ours bears on a service of ours, the conflict is stated on the page.
semanticeconomy.org and spxi.dev are co-constitutive surfaces of one practice: spxi.dev carries the SPXI protocol and its methods; this site carries the economics the protocol answers and the services that carry it out. Each is read with the other.
The knowledge graph is a medium. The entity you build in it is the sign that holds your meaning when you are not in the room.
What do you want to build?
— Rex Fraction · 06.SEI.ARCHIVE
∮ Build it
The knowledge graph is being written right now. What do you want it to say?
What do you want to build for in the knowledge graph?
The knowledge graph currently has no canonical definition for — or one composed from whatever the AI could find. You can build something better. Something durable. Something yours.
An SPXI baseline audit maps what exists now — what the AI composes, what it burns, what's missing — and produces the blueprint for building. Five business days. Fixed scope.
Hugging Face dataset(the archive as Parquet: deposits, captures, reception, predictions — rebuilt from the registry)
dataset query endpoint(no auth, no client; config=citations for the edge list. /search needs an index that rebuilds after every deposit and 500s while it does — use /rows)