Know Where Every
Number Came From
Prometheux records lineage as it computes, not by parsing code afterwards. Follow every result back through the rules, concepts and sources it depends on, across every connected source.
Proven in production
Arrows Between
Tables Aren't Proof
Most lineage tools show that table B was built from table A. Auditors ask a harder question: which rows, through which logic, produced this figure?
Lineage Recorded
as It Runs
When the engine reasons over your data it builds the chase graph: each derived fact alongside the facts and rule that produced it. Turn provenance on and every output traces back to source data.
Fact, Sources, Rule
Each provenance record holds the derived fact, the source facts that produced it and the rule that was applied.
Recursion, Unrolled
Supply chains and multi-hop paths are traced one derivation at a time, however deep they go.
CSV or Neo4j
Write provenance to CSV, or into Neo4j as CHASE_NODE nodes and DERIVED_BY edges you can query with Cypher.
Explanations on Top
@explain turns provenance into JSON and text explanations, written to CSV, PostgreSQL or the console.
Answer audit questions down to the row. Prove how a figure was computed, not just where it was stored.
See the Whole
Chain at a Glance
Above the rows sit your concepts and rules. The lineage canvas in the Ontology Workspace shows how concepts depend on each other and on the data sources connected to them.
Concept Dependencies
Follow a result from the concept that produced it, through upstream concepts, to the connected data sources.
Across Sources, In Place
Concepts read warehouses, lakes and databases where they are, so lineage spans systems without copying data.
Understand impact before you change a rule. One lineage view across your whole estate.
From Source Row
to Audit Trail
Connect Your Sources
Connect warehouses, lakes and databases to Prometheux. Data stays where it is.
Define the Logic
Write transformations and calculations as Vadalog rules in named concepts.
Run With Provenance On
The engine records every derivation as it reasons, into CSV or Neo4j.
Trace Any Result
Follow a figure back through rules and facts to the original source rows.
Lineage Where It
Is Non-Negotiable
Patient Journeys
Lab results, procedures and medications joined into end-to-end patient journeys, with lineage down to each individual patient ID.
Read the story →Regulatory Reporting
Banca Sella's credit risk explanations come with end-to-end data lineage, cutting a regulatory flow from 5 hours to under a minute.
Read the case study →Aggregates You Can Unpick
Trace a reported total back through each calculation step to every contributing row.
Root-Cause Analysis
When a figure looks wrong, provenance shows which rule and which input rows produced it.
“The ability to query heterogeneous databases through a flexible ontology makes it possible to infer deep knowledge and automate complex tasks, delivering trustworthy insights that far exceed the capabilities of traditional systems.”
Anna FavaFormer Senior Manager, Data & AI · Dompé
“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.”
Financial institution
Questions,
Answered
What level of lineage does Prometheux provide?
Follow every result back through the concepts and sources it depends on. The chase graph records each derived fact together with the source facts and the rule that produced it.
How is data lineage different from AI lineage?
Data lineage covers deterministic results: which rules and source rows derived a value. AI lineage extends the trail through generative steps, showing which LLM step, prompt, model and context notes shaped an AI answer.
Does this replace our data catalogue?
No. Catalogues map lineage between tables and systems across your estate. Prometheux adds what happens inside the logic: for every result it computes, the rules and source facts behind it. Provenance is written to CSV or Neo4j, so you can take it wherever you need it.
Can I check the impact of a change before I make it?
Yes, at concept level. The lineage canvas in the Ontology Workspace shows how concepts depend on each other and on their connected sources, so you can see what sits downstream of a rule before you change it. For root cause, provenance shows which rule and which source facts produced a figure.
Where is lineage stored?
Provenance is written to CSV, or to Neo4j as CHASE_NODE nodes connected by DERIVED_BY edges that you can query with Cypher.
Related Capabilities
Trace Your Numbers
to the Row
Bring a figure your auditors question. We'll show you the rules and rows behind it.
