Memory schemas

Schema reference for Curia’s knowledge graph — node types, edge types, decay classes, sensitivity levels, and temporal metadata.

Curia’s knowledge graph stores all long-term memory in Postgres as typed nodes and edges with temporal metadata. This page documents the data model.

Node types

Every node in the knowledge graph has a type field from this fixed set:

Type	Description	Examples
person	An individual	CEO, board members, contacts
organization	A company, team, or institution	Acme Corp, Finance team
project	A project or initiative	Q3 Fundraising, Website Redesign
decision	A decision that was made	”Hired Alice as CFO”, “Moved to weekly standups”
event	A meeting, conference, or occurrence	Board meeting, Series A close
concept	An abstract idea or topic	Machine learning, Compliance policy
fact	A specific piece of information about an entity	”Prefers morning meetings”, “Lives in Toronto”

Edge types

Edges represent typed relationships between nodes:

Category	Types	Description
Structural	works_on, decided, attended, relates_to, belongs_to, authored, mentioned_in	General relationships between entities
Personal	spouse, parent, child, sibling	Family relationships
Professional	reports_to, manages, collaborates_with, advises, represents	Organizational hierarchy and work relationships
Organizational	member_of, founded, invested_in	Membership and investment relationships

relates_to is the generic fallback — use it when no specific type fits.

Temporal metadata

Every node and edge carries temporal metadata:

interface TemporalMetadata {
  createdAt: Date;         // when first recorded
  lastConfirmedAt: Date;   // last time evidence reinforced this
  confidence: number;      // 0–1 scale, decays over time
  decayClass: DecayClass;  // how fast confidence decays
  source: string;          // provenance: "agent:coordinator/task:abc/channel:cli"
}

Decay classes

Class	Half-life	Behavior	Examples
permanent	Never	Confidence never decreases	Legal name, date of birth
slow_decay	~180 days	Halves every 6 months	Employer, city of residence
fast_decay	~21 days	Halves every 3 weeks	Coffee preference, current project focus

Confidence decays using a half-life model: new_confidence = current × 0.5^(days / half_life). See Dreaming for how the decay pass works.

Sensitivity levels

Every node has a sensitivity classification assigned at creation time:

Level	Description	Export behavior
public	No restrictions	Freely exportable
internal	Default for most facts	Exportable with standard authorization
confidential	Sensitive business information	Requires elevated authorization
restricted	Highest sensitivity	Bulk export blocked

Sensitivity is immutable after creation — changing it requires a manual data migration. Skills can provide an explicit sensitivity override when storing facts, or let the system auto-classify using keyword analysis.

Node schema

interface KgNode {
  id: string;                          // UUID
  type: NodeType;                      // person, organization, project, etc.
  label: string;                       // display name
  properties: Record<string, unknown>; // arbitrary key-value data
  embedding?: number[];                // VECTOR(1536) for semantic search
  temporal: TemporalMetadata;
  sensitivity: Sensitivity;
}

Edge schema

interface KgEdge {
  id: string;                          // UUID
  sourceNodeId: string;                // UUID of the source node
  targetNodeId: string;                // UUID of the target node
  type: EdgeType;                      // works_on, reports_to, etc.
  properties: Record<string, unknown>; // arbitrary key-value data
  temporal: TemporalMetadata;
}

Storing facts

When storing a new fact via the memory-store skill:

interface StoreFactOptions {
  entityNodeId: string;            // the entity this fact is about
  label: string;                   // the fact text
  properties?: Record<string, unknown>;
  confidence?: number;             // default: 0.7
  decayClass?: DecayClass;         // default: "slow_decay"
  source: string;                  // provenance string
  sensitivity?: Sensitivity;       // override auto-classification
  sensitivityCategory?: string;    // hint for auto-classifier
}

Validation

Before creating a new fact, the memory validator checks:

Check	Outcome
Deduplication (cosine similarity > 0.92)	Updates lastConfirmedAt and merges properties instead of creating a duplicate
Contradiction (conflicting fact exists)	Higher confidence wins; lower is archived with previous values preserved
Equal confidence conflict	Flagged for CEO review via alert channel
Write rate limit (50 per task)	Excess writes dropped with audit log warning

Memory and the knowledge graph

Conceptual overview of the four memory tiers and how they work.

Dreaming

How the decay pass processes temporal metadata overnight.