Technology & engineering
Operate serious graph-based work without assembling another fragile stack.
Legra brings graph storage, SPARQL, Cypher, GQL, RDFS, OWL-RL, Datalog, validation, branching, task execution, provenance, content addressing, encryption, and peer-to-peer communication into one operating model. Technology teams can run it locally, across organizational infrastructure, or through Operators without giving one vendor authority over every workspace.
What changes
Use RDF and labelled-property-graph interfaces over one graph infrastructure.
Bind queries, validation, inference, search, task execution, and generated artefacts to versioned workspace state.
Keep workspace data encrypted and distribute custody separately from the authority to interpret it.
Give humans and agents the same explicit stories, permissions, budgets, review paths, and evidence model.
Connect existing code and services through APIs and Story implementations instead of rewriting everything as graph data.
Deploy nodes on your own hardware and use network services when activity crosses economic or organizational boundaries.
Inspect the security, architecture, and operational trade-offs directly rather than relying on a generic platform claim.
Infrastructure lens
Run one journaled path from admission and metering to settlement.
Every priced operation follows the same typed lifecycle. Workspace authority and a token budget admit the activity; Hosts, Custodians, Workers, Licensors, and external providers expose measured contributions; task and activity journals preserve the evidence used for settlement.
Which principal authorized the operation and which rights applied?
Which node or provider consumed each measurable resource?
Which task, input, version, derivation, and result support the charge?
Which atomic settlement entries credit every valid counterparty?
Zero-priced local work and positive cross-party exchange use one accounting path, while evidence remains distributed with the workspace activity instead of one opaque billing database.
Economic modelOne graph, several interfaces
RDF semantics and LPG ergonomics do not need separate islands. Legra is designed so SPARQL, Cypher, GQL, Datalog reasoning, validation, vector search, and full-text search participate in one branch-sensitive knowledge context.
ExploreAuthority and custody are different
Nodes that retain encrypted blocks do not automatically gain the right or keys to read them. Identity, workspace authority, policy, signed history, and task evidence remain explicit across the network.
ExploreAI work is ordinary governed work
Agents act through the same Use Cases, Stories, task trees, branches, budgets, tools, and review mechanisms as other participants. Model providers remain replaceable; the durable context and evidence stay in the workspace.
ExploreContinue exploring