Products
Kinetic Graph
Links between entities as they change. Cause you can walk, not two numbers that moved together.
People, processes, and payments are nodes
The Kinetic Graph is a network. Each entity — a person, a process, a transaction — is a node. The links between them are edges, and those edges keep changing.
A warehouse stores a snapshot. The Kinetic Graph stores cause: who affects whom, how a change travels, and which links get stronger or weaker.
Kinetic means the graph moves. Links update live. New ones appear. Old ones fade. You see what is doing work, not just what exists.
Node ANode B
Node CNode D
Node ENode F
Graph Database Structure
Inputless DB stores 22 node types and 30+ relationship types.
Core Entities
User
User identifier and metadata
Session
Session context and duration
Event
Behavioral events and interactions
CognitiveState
Intent state and confidence
Pattern
Detected behavioral patterns
Anomaly
Unusual behavior detection
Relationships
HAS_SESSION
User → Session
CONTAINS
Session → Event
ENRICHED_BY
Event → CognitiveState
FORMS
Event → Pattern
TRIGGERS
Event → Anomaly
GENERATES
Pattern → Insight
Graph RAG
Ask in language. Graph RAG walks Inputless DB and a language model reads the subgraph.
How it works
01
Ask in language
A question against the graph you already have
02
Walk the graph
Relevant nodes and links in Inputless DB
03
Read the subgraph
A language model writes from those nodes, not from the open web
04
Answer with a trail
The call, the evidence nodes, and a confidence figure
Use Cases
"Which sequences of events precede a conversion"
"Which sequences precede churn"
"Which events actually move the outcome, not just sit next to it"
"Where anomalies cluster across a cohort"
"How strong a link is between two entities"