Deep memory as a service. A specialized database for entity-relationship graphs with neural-inspired storage. Your organization's knowledge woven into a living graph — every entity known, every relationship traced.
NERD stores your world as entities — people, documents, concepts, events, code. Every one uniquely identified, richly typed, and deeply connected.
Team members, clients, users, stakeholders. Each linked to every document they authored, every decision they influenced.
entity_type: personMarkdown, PDFs, code files, meeting notes, specifications. Ingested, indexed, and woven into the knowledge graph.
entity_type: documentAbstract entities — design patterns, architectural decisions, principles. Traced across every project they touch.
entity_type: conceptDeployments, releases, meetings, deadlines. Temporal anchors that connect entities through time.
entity_type: eventFunctions, modules, services, packages. Every unit of code linked to its documentation, tests, and dependents.
entity_type: code_unitTeams, departments, companies, projects. Hierarchical entities that group and contextualize everything below.
entity_type: organizationEntities alone are static. NERD's power is in the edges — typed, directed relationships that turn a collection into a living knowledge web.
Documents belong to projects. Code belongs to modules. People belong to teams. The fundamental containment relationship that gives your graph hierarchy.
entity —[belongs_to]→ entity
When a document mentions a concept, cites a decision, or links to another entity. The backlink trail that makes knowledge discoverable across silos.
entity —[references]→ entity
Specifications derived from requirements. Implementations derived from designs. Tests derived from specs. The provenance chain of every artifact.
entity —[derived_from]→ entity
Undirected semantic affinity. Two entities that share context, co-occur frequently, or the graph has learned are connected through use patterns.
entity —[related_to]— entity
Start with any entity — a person, document, concept
▼ 3 edgesFollow typed edges — belongs_to, references, derived
▼ 12 nodesDiscover the full subgraph — connected entities, clusters
▼ depth-firstSurface patterns — who influences what, which ideas connect
◉ 47 connectionsNERD is engineered for depth. Every layer designed for entity-relationship storage, traversal, and retrieval at production scale.
From a single team to an entire organization. NERD grows with your knowledge graph.
For small teams. Self-hosted NERD with up to 50K entities.
For growing orgs. Up to 500K entities with advanced graph features.
For organizations with millions of entities. Managed or self-hosted.
Every entity in your organization is waiting to be discovered. NERD weaves them into a living graph — searchable, traversable, and always growing.