Legra Features
Legra is not just another graph database. It is infrastructure for sovereign data — where every dataset you create is yours to control, share, version, and eventually monetize, without handing ownership to a platform or a middleman.
The features below are the building blocks that make this possible.
Branch your data like code
Isolate changes on a branch. Compare alternatives. Merge when ready. Roll back if not.
Encrypted like Signal
Your data is encrypted before it leaves your machine. Only workspace members hold the keys.
No single server
Run your own nodes. Work offline and sync later. Popular data replicates as people use it.
Three languages, one graph
One unified model for SPARQL, Cypher, and GQL. Any query language, any result format.
A graph that checks itself
Define rules once. Every query — in any language — includes inferred facts. Validation catches bad data.
Full-text search inside the graph
Ranked, language-aware text search with no separate search engine — versioned, encrypted, and queryable in one statement with graph traversal.
AI-ready without a second stack
Graph-derived similarity search with no external vector database, no middleware, no stitching.
Validate as you build
Set rules for what good data looks like. A staging branch warns instead of blocking, so you can finish, then merge into production knowing it passes.
See what's happening
Live Tasks and Journal panels show every operation in your workspace — progress, where it ran, and what failed — without leaving Studio.
Events with provenance
Record domain changes as queryable graph events linked to commits, tasks, actors, and causes.
Give agents a job, not credentials
Bind an agent to one persona, one Use Case tree, tested Stories, a task branch, and a token budget.
Adopt reusable business capability
Discover Apps, Data Products, automations, transforms, and services as governed Use Case packages, then adopt them through a reviewable branch.
Settle every contribution
Compare whole-token offers, authorize funding, prove delivery, reward contributors, and settle every directed claim.
Built for AI agents
Most AI agent frameworks treat the knowledge layer as a retrieval problem — embed some documents, search by similarity, hope the context is good enough. Legra gives agents something better: a structured world model they can query, reason over, and write to — with built-in guardrails.
Give an agent a scoped workspace with only the data its domain requires. The agent queries the graph for real relationships — not text fragments. It writes new knowledge to a branch, where it can create and experiment freely without touching production data. A human reviews the branch, sees exactly what the agent produced, and merges or discards — the same review workflow developers use for code, applied to data. Every agent action can be tied to a signed commit, task provenance, and domain events that record what changed.
The features on this page are not independent capabilities. Together, they form the infrastructure that makes this possible: encryption controls agent access, workspaces define agent scope, branching provides the sandbox, merge provides human oversight, and the audit trail and event provenance are automatic.
See Agent Guardrails for the full isolation hierarchy, from persona-bound Story access to deliberately weaker raw statement modes.
See Use Case Exchange for the universal package model behind Apps, Data Products, transforms, services, and other reusable business capabilities.
See Token Accounting and Settlement for the constant unit, competitive offers, sponsored activities, and directed settlement graph.
Your data is yours
Legra workspaces are sovereign data assets — encrypted, versioned, and controlled by you. Share what you choose, with whom you choose, on your terms. No platform lock-in. No middleman. When another party uses positively priced data or capability you publish, the resulting token settlement can compensate you directly.
Where to start
If you are evaluating Legra for broader fit, begin with Three Query Languages, then Branching and Merge Workflows, then End-to-End Encryption. If agents are part of the deployment, continue with Agent Guardrails.
For procurement, standards, governance, and risk reviews, see Frameworks & Standards.
For hands-on usage, move into the Guides or the Legra Documentation index.