API — Conceptual Hypergraph

Documentation status: reference — see Maturity and evidence.

This section documents the public conceptual-hypergraph model used by logiCells.

The conceptual hypergraph is the shared structure behind conceptual knowledge, explicit semantic relations, H-Logic, symbolic/hybrid reasoning, persistent conceptual state and numerical projections.

Main areas

• Hypergraph access and programming
• Type constraints
• Special dimensions
• Indexing
• H-Logic
• Embeddings and conceptual search
• Tensor projections
• Persistent hypergraph
• Durability and WAL

Hypergraph and vector search

The current engine keeps the symbolic and numerical layers connected without collapsing them. A conceptual value can be associated with an embedding vector, indexed through native HNSW retrieval, then recovered as a conceptual candidate. Symbolic Kind constraints can reject numerically close but semantically incompatible candidates.

That separation is intentional: numerical similarity proposes and ranks; conceptual structure remains authoritative for identity and validity.

Public-contract rule

Pages in this section describe conceptual and public Runtime behavior. Private storage layouts, implementation-language records and internal engine classes are not public API contracts.