Case Study
How Sella Delivers Scalable Credit Risk Explanations with Prometheux
Read More
Capabilities / Graph & Complex Data Processing

See the Logic.
Trace the Data.

Prometheux draws your business logic as a live lineage graph of concepts, rules and sources, so anyone can see how a number is built, whether each step ran and which rows it came from, without writing a query.

Proven in production

Why graphs get ignored

Hairballs Hide.
Lineage Explains.

A graph with every edge on screen at once tells nobody anything, however good the canvas. Prometheux shows the structure that matters: what each result depends on, whether it ran, and the rows behind it.

Typical graph view
Graph visualization with Prometheux
Hundreds of edges drawn at random
A dependency graph from sources to concepts to results
Pretty diagram, disconnected from runs
Each concept shows a live status: running, success, error or skipped
No way to see what feeds a KPI
Click a concept to highlight its upstream dependencies
Source data is a separate tool away
Open any source to see its schema and a data preview
Lineage of results lost after the run
The chase graph can be explored in Neo4j or exported as CSV
Analysts rebuild the story for each review
The canvas is drawn from the rules themselves, so it updates as the logic changes
The lineage view

Your Logic,
Drawn Live

The ontology workspace centres on a lineage canvas. Concepts are cards, data sources are nodes and edges show how data flows, with an Explorer tree and panels for concepts and data alongside.

Status at a Glance

Every concept card carries a status dot, so failed or skipped steps stand out in the graph.

Upstream in One Click

Select a concept to open it in the editor and highlight everything it depends on.

Sources You Can Open

Data source nodes show connection details, schema and a searchable, paginated data preview.

Built for Big Graphs

Collapse or expand all cards, zoom and fit, and navigate by the Explorer tree with type indicators.

What this means for you ⟶

See your logic drawn

Onboard new team members to the model faster. Spot the broken step before the report goes out. Show reviewers exactly what feeds a number.

Result lineage

From Answer
Back to Row

Behind every result is a chase graph: the rules and facts that produced it. Export it as a table or materialise it to Neo4j to navigate large lineage graphs interactively.

Fact-Level Provenance

Each derived fact records the facts it came from and the rule that generated it.

Explore at Scale

Materialise lineage to Neo4j for interactive navigation of large provenance graphs.

Query the Graph Too

Use Cypher MATCH patterns over the same tables to pull exactly the subgraph you want to see.

What this means for you ⟶

Talk to an engineer

Answer "where did this number come from?" with a picture. Give auditors lineage they can walk through.

How it works

From Model to
Picture to Proof

01

Define Your Rules

Connect sources and define concepts in Vadalog, SQL, Cypher or Python.

02

See the Lineage

The workspace draws the dependency graph automatically as concepts are added.

03

Run and Watch

Status dots update per concept as runs succeed, fail or skip.

04

Trace a Result

Follow any result through its chase graph back to the rules and rows that produced it.

Who uses it

One Picture for
Every Stakeholder

Data teams

Debug Pipelines Visually

See which concept failed and what depends on it, without reading logs.

Risk & audit

KPI Lineage

Show the sources and rules behind a regulatory figure, step by step.

Business

Understand the Model

Give domain experts a map of the concepts that define their business.

AI teams

Explain Agent Answers

Trace an agent's answer through the concepts and data it relied on.

Graphs Big Enough to Need a Map

“With Prometheux, we achieved a speed-up of over 250% on critical tasks involving complex reasoning over knowledge graphs with hundreds of millions of nodes and billions of edges.”
HR Head of R&D
Financial institution
“Prometheux technology powers highly scalable and explainable reasoning over some of the world's largest knowledge graphs.”
NV NVIDIA Technical Blog
Accelerating Neurosymbolic AI with RAPIDS and Vadalog Parallel
FAQ

Questions,
Answered

What does the graph show?

The lineage view shows your concepts and rules as a dependency graph: concept cards, data source nodes and the edges between them, with each concept's run status.

Can I see the data behind a node?

Yes. Opening a data source node shows its connection details and schema, plus a paginated, searchable preview of the data.

How do I trace a single result?

With provenance on, Prometheux records a chase graph of the rules and facts behind every derived result. You can export it as CSV or materialise it to Neo4j to explore it interactively.

Is this for exploring instance data, like an investigation graph?

The workspace focuses on the model and its lineage. To look at instance data, pull the subgraph you need with a Cypher pattern, or materialise result lineage to Neo4j and explore it there.

Do I need a graph database to use it?

No. The lineage view is built into Prometheux. Neo4j is optional, for interactive exploration of large result lineage.

Who is it for?

Anyone who needs to understand a result: data teams debugging a run, risk and audit reviewing a figure, and domain experts reading the model. The lineage view is drawn automatically from the concepts, so nobody has to write a query to see it.

See How Your
Answers Are Made

Bring your business rules or a report you need to explain. We'll show its lineage end to end.

We use some essential cookies to make this site work. By clicking "Accept All" you agree to the storing of cookies on your device to enhance site navigation, analyse site usage and assist our marketing efforts. For more detailed information see our Cookie Notice.

Accept AllDecline