Follow Every Link.
At Any Depth.
Ask who is exposed, what depends on what and how two entities connect, however many hops apart. Prometheux follows relationships across the data you already have and returns the route behind every answer.
Proven in production
One Hop Is Easy.
Ten Is the Point.
A join answers the first level. The questions that matter, like every plant a supplier feeds or every service an outage hits, don't come with a known number of hops. Recursive CTEs often cap the depth; graph databases want a copy of the data.
Every Path,
Shortest Path
Traversal functions take any relation of edges and return paths as rows. Choose a source, a target, a set of pairs or a maximum depth, and keep the visited nodes so every answer shows its route.
Reachability
Reachability lists everything a node can reach, and everything that can reach it: its full downstream footprint.
Paths With Their Route
All paths, with the nodes visited, from one source, to one target or between chosen pairs.
Shortest Routes
Shortest paths from one source or between all pairs, over weighted edges such as cost, distance or time.
Hop by Hop
A recursive rule with a minimum counts how many hops each node sits from the start, so you can see first, second and nth-degree links.
See the full chain, not just the first level. Show the route behind every connection. Answer new what-if questions without new pipelines.
Multi-Hop in
Plain Cypher
Analysts can describe a traversal as a Cypher pattern, including variable-length hops and shortestPath. Prometheux runs it over CSV, Parquet, Iceberg, SQL databases or Neo4j, without loading a graph store first.
Variable-Length Hops
Express "one to six hops away" as a single pattern instead of six joins.
Filter Along the Way
WHERE clauses with comparisons, regex and string predicates filter the pattern.
Edges From Anywhere
Connect suppliers from Postgres and shipments from Parquet; the pattern doesn't care where each edge lives.
Paths Into Rules
Traversal results become predicates that downstream rules aggregate, score or turn into alerts.
Let analysts write traversals they can read. Run them on today's data, not last night's copy. Turn paths straight into decisions.
From Question
To Full Chain
Map the Relationships
Connect the tables that hold supply, payment, reporting or dependency links. Data stays where it is.
Pick a Start or End
Choose the source, target or pairs you care about, and an optional depth limit.
Traverse and Explain
Get every reachable node and the route to it, or the shortest route, as rows you can inspect.
Questions With
No Fixed Depth
Fraud Rings
Trace money between accounts through every intermediary, however many hops deep, and show the route.
Finance & banking →Supplier Blast Radius
If this supplier fails, what breaks next? Follow supply links to every component, product and customer affected.
Manufacturing →Biomedical Pathways
Connect compounds to diseases through genes and targets across a knowledge graph of 4.7M data points.
Pharma →Dependency Mapping
Find every service downstream of a failing site, and the shortest alternative route around it.
Telecom →“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
“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
Questions,
Answered
What is multi-hop graph traversal?
Following relationships across several steps: from a supplier to every product it feeds, or from a failing server to every service that depends on it. Prometheux handles paths of any length, or up to a depth you set.
Can I see the path, not just the result?
Yes. With visited=true, path functions return the nodes along each route, so every connection comes with its explanation.
Which traversal functions are available?
Reachability, all paths and shortest paths (from one source or between all pairs), plus Cypher multi-hop and variable-length patterns.
Do I need a graph database for multi-hop queries?
No. Traversal functions and Cypher patterns run over relational tables and files in place. Neo4j is supported as a source if you already have one.
Do the relationships need to be in one database?
No. Edges can come from different sources, such as a SQL database and Parquet files, and be combined into one traversal without moving the data.
Will I get the same answer every time?
Yes. Traversal functions and rules are deterministic: the same data and the same definitions return the same paths, and each path carries the nodes it visited.
How large a graph can it traverse?
Execution is parallel and distributed. A financial institution reported a speed-up of over 250% on reasoning over graphs with hundreds of millions of nodes and billions of edges.
Related Capabilities
Trace Your Longest
Chain in Minutes
Bring the relationships you can't follow today. We'll traverse them end to end on your own data.
