Reusable Cells
WordCells and ActionCells retain persistent identities and reusable lexical or action knowledge without sentence backlinks or copied facts.
Transparent language architecture
Vireon: cell-based semantic language intelligence
Vireon stores words and actions as reusable Cells, while source-ordered SentenceCells preserve the relationships, context and evidence needed to answer questions.
Architecture updated: July 2026
SentenceCell evidence
SentenceCells carry source order, token identities, grammatical roles and compact sentence-local connections. They are the primary evidence for learned relationships.
Deterministic retrieval
Saved SentenceCells can be re-read in their original order to rebuild discourse context and answer questions independently of the order in which those questions are asked.
Current stable foundation: V1.0
What V1.0 establishes
V1.0 connects information across multiple SentenceCells and retrieves entity timelines, problems, actions, before-and-after events, repair processes, outcomes and participants from ordered evidence.
Read the technical architectureNext semantic layer
The nodes remain the Cells. Each Cell will also connect through structured fields:
key → class → subcategory → category → subdomain → domain
Sentences establish contextual meaning first; category fields then add a second route between related Cells.
Explore the semantic modelPublic demonstration plan: the first web demo will use a prebuilt, read-only population of several thousand Cells. Visitors will ask questions using existing Cell knowledge and logic; the public interface will not create or alter Cells.
Projects
Vireon
A meaning-first language intelligence project built from reusable Cells, sentence evidence and transparent retrieval paths.
View Vireon project informationViaLens.live
An interactive visual project combining photography, wildlife and nature identification, educational notes, games and quizzes.
Visit ViaLens.live