Knowledge graphs, engineered at scale.

Taxonomies have reached their limits.

JCAT Labs is the infrastructure layer for enterprise knowledge. We model, align, govern and serve taxonomies, ontologies and knowledge graphs across billions of entities, then plug them into every system that needs to reason over your data: search, analytics, RAG and AI agents.

The engine is domain-agnostic. It structures a gallery, library, archive or museum collection as readily as a defence or intelligence vocabulary. Open source and free forever; we make the hard part, curation and managed infrastructure at scale, simple.

The engine is open. The infrastructure is the product.

£0

Dump the mess. We sort it.

Documentation, guidelines, spreadsheets, legacy schemas, raw exports: hand JCAT whatever you have and it sorts all of it into one governed taxonomy or ontology. Standing up a triple store takes an afternoon; keeping it correct, aligned and governed as it grows to enterprise scale is the hard part, and it is where teams stall. The engine that models, validates, aligns and serves is open-source and free forever. You pay only when you want it curated and hosted for you.

Everything between raw data and a graph you can trust.

Four capabilities, one source of truth for your whole organisation.

Model & ingest

Author taxonomies and ontologies, or import what you already have: RDF, OWL, SKOS, spreadsheets, legacy schemas. Mappings are generated and validated against your shapes, not hand-typed.

Align & reconcile

Merge overlapping vocabularies across teams, vendors and acquisitions. Candidate matches are scored, adjudicated and explained, so reconciliation is auditable rather than a black box.

Govern & version

Every change is diffed, reviewed and versioned with full lineage. Roll back, branch, and prove exactly what your ontology looked like on any date, for any audit or regulator.

Serve at scale

One API and SPARQL endpoint for the entire organisation. Sub-100ms queries over billion-entity graphs, feeding search, analytics, RAG pipelines and agents from a single source of truth.

Try it

Paste a list. Get a graph.

Drop in a list of terms and jcat-mini renders it as a SKOS taxonomy or an OWL ontology, then exports the Turtle. This is the open model running live on Hugging Face. Start simple here, then bring the mess: dump your documentation, guidelines, spreadsheets and raw data, and the full product sorts all of it into one governed graph.

The hook

Free. Forever. Not a trial.

The JCAT engine is a lightweight, open-source Python core. The thing that models, validates and serves your knowledge graph is yours to run, free, with no seat limits and no expiry. We never paywall the engine.

›Open core. Run it locally, in your VPC, or on a laptop. No lock-in, no rug-pull.
›We charge for the hard work. Curation, managed hosting, alignment at scale, SLAs and support. The work, not the software.
›Graduate when you want. Start free, move curated and hosted only when scale, compliance or uptime demand it.

# pip install jcat, that is the whole engine
from jcat import Graph

g = Graph.load("catalog.ttl")
g.align("vendor_taxonomy.csv")
g.validate(shapes="shapes.ttl")

# serve it: one line, billions of entities
g.serve(port=8080)

# > live . <100ms . SPARQL + REST

Three models. One free and open.

The same engine in three sizes. jcat-mini is open and free forever on Hugging Face. Step up to managed curation and infrastructure when scale, compliance or uptime demand it.

jcat-mini · Open

£0 / forever