Stable identifiers
Concept IDs remain addressable as embedding models, vector dimensions, and retrieval techniques change.
About the project
Semantic identity should outlive the model that found it.
Embedded Semantics is designed around a simple separation: machine-learned vectors estimate semantic proximity; a registry records stable meaning.
Multilingual evidence
Language expressions attach to concepts with locale, review status, equivalence strength, and provenance.
Explicit uncertainty
Similarity scores are evidence. Resolver thresholds and abstention keep ambiguity visible rather than hiding it.
Composable meaning
Long-term representations can combine concepts, relations, and semantic residue rather than forcing a sentence into one opaque vector.
Design boundary
What the registry does not claim
It does not claim that every culture partitions meaning identically, that a single vector captures every nuance, or that translation is lossless. The system is designed to preserve unresolved distinctions as metadata and evidence.