Knowledge Graph API Reference

All endpoints live under /api/v1/rdf unless noted. Every one returns 503 when RDF is disabled. Authenticate with a bearer token:

export COLLATE_HOST=<your-collate-instance-url>
export TOKEN=<your-personal-access-token>

Most knowledge-graph endpoints are admin-only. Graph node hydration resolves entity details without re-applying the caller’s per-entity view authorization, so opening these to non-admins would leak entities a user cannot otherwise see. The glossary-scoped ontology endpoints on /api/v1/glossaries are the non-admin path.

Status

Endpoint	Access	Returns
`GET /status`	Authenticated	`baseUri`, `enabled`, `storageType`, `inference` (`enabled`, `defaultLevel`, `availableLevels`), `projectionState` (`DISABLED` / `READY` / `REBUILDING` / `DEGRADED`), `askCollateEnabled`

SPARQL

Endpoint	Access	Notes
`GET /sparql`	Admin	`query`, `format` (`json`, `xml`, `csv`, `tsv`), `inference` (`none`, `rdfs`, `owl`, `custom`)
`POST /sparql`	Admin	JSON body `{ "query": "...", "format": "...", "inference": "..." }`
`POST /sparql/update`	Admin	SPARQL UPDATE. See the warning below.
`GET /queries/saved`	Authenticated	The caller’s private saved queries.
`PUT /queries/saved`	Authenticated	Replace the caller’s saved queries.
`GET /queries/templates`	Authenticated	Installation-wide shared query templates. Writes are admin-only, via system settings.

POST /sparql/update writes directly to a derived index. Anything you write is discarded on the next reindex. Use the entity APIs to change your catalog.

Entities and Graph

Endpoint	Access	Notes
`GET /entity/{entityType}/{id}`	Admin	The entity as RDF. `format`: `jsonld` (default), `turtle`, `rdfxml`, `ntriples`.
`GET /entity/{entityType}/{id}/diff`	Admin	Diff two RDF versions of an entity.
`GET /graph/explore`	Admin	`entityId`, `entityType`, `depth` (default 2), `entityTypes`, `relationshipTypes` (comma-separated filters).
`GET /graph/explore/export`	Admin	Same parameters plus `format`: `turtle` (default) or `jsonld`.
`GET /ontology`	Authenticated	Download the canonical OpenMetadata ontology.
`GET /glossary/graph`	Authenticated	Glossary term relationship graph. See parameters below.
`GET /glossary/{id}/export`	Authenticated	Export a glossary as an ontology.
`GET /inference/lineage/{entityId}`	Admin	Full lineage with inference applied. Requires `entityType`.
`GET /debug/glossary-relations`	Admin	Diagnostic view of glossary relations in RDF.

`GET /glossary/graph` Parameters

Parameter	Default	Purpose
`glossaryId`	—	Scope primary terms to one glossary.
`glossaryTermId`	—	Scope to one term and its direct incoming/outgoing relations.
`relationTypes`	—	Comma-separated relation types to include.
`limit`	`500`	Maximum terms returned.
`offset`	`0`	Pagination offset.
`includeIsolated`	`true`	Include terms with no relations.

Passing both glossaryId and glossaryTermId requires the term to belong to that glossary while still returning direct cross-glossary neighbors.

Validation

Endpoint	Access	Notes
`POST /validate`	Admin	SHACL validation. `entityUri` to scope to one entity (omit for the whole dataset), `format`: `turtle` (default) or `jsonld`. Response carries an `OM-SHACL-Conforms` header.

Inference Rules

Endpoint	Access	Notes
`GET /rules`	Admin	Rules with `dirty`, `graphUri`, `tripleCount`, `lastMaterializedAt`, `lastError`, `systemRule`.
`GET /rules/{name}`	Admin	One rule.
`PUT /rules/{name}`	Admin	Create or update. Marks the rule dirty.
`DELETE /rules/{name}`	Admin	Delete a custom rule and drop its graph. System rules cannot be deleted.
`POST /rules/validate`	Admin	Validate a rule body without saving.
`POST /rules/materialize`	Admin	Materialize now.

Custom Ontology Extensions

Endpoint	Access
`GET /ontology/extensions`	Admin
`GET /ontology/extensions/{name}`	Admin
`PUT /ontology/extensions/{name}`	Admin
`DELETE /ontology/extensions/{name}`	Admin
`POST /ontology/extensions/validate`	Admin

Custom classes and properties must be in https://open-metadata.org/ontology-extension/.

Insights

Endpoint	Parameters
`GET /insights/important`	`entityType` (required), `window` (`daily`/`weekly`/`monthly`), `limit` (1–100, default 20)
`POST /insights/recompute-centrality`	`entityType` (required)
`POST /insights/recompute-communities`	`entityType` (required), `graphType` (`lineage`/`tagCoOccurrence`)
`GET /insights/communities`	`entityType` (required), `graphType`
`GET /insights/path`	`from`, `to` (required URIs), `direction` (`upstream`/`downstream`/`both`), `maxHops` (1–25, default 6)
`GET /insights/recommendations`	`entityUri` (required), `limit` (1–50, default 10)
`GET /insights/tag-cooccurrence`	`minCount` (default 2), `limit` (1–100, default 20)
`GET /insights/glossary-reach`	`minDomains` (default 2), `limit` (1–100, default 20)
`GET /insights/tag-popularity`	`limit` (1–100, default 20)

All admin-only.

Semantic Search

Endpoint	Notes
`GET /search/semantic`	Vector search over the graph. Requires OpenSearch.
`GET /search/similar/{entityType}/{id}`	Similar entities. Requires OpenSearch.
`GET /search/recommendations/{userId}`	Per-user recommendations. Requires OpenSearch.

SQL → SPARQL

Endpoint	Notes
`POST /sql/translate`	Translate SQL to SPARQL.
`POST /sql/query`	Translate and execute.

Linked Open Data

Endpoint	Notes
`GET /api/v1/lod/entity/{entityType}/{id}`	Authenticated, content-negotiated redirect for a minted IRI. Requires `RDF_DEREFERENCEABLE_IRIS=true`.

Glossary Ontology Endpoints

These live under /api/v1/glossaries and, importantly, work with RDF storage disabled, because they read the database-primary model.

Endpoint	Access	Notes
`PUT /api/v1/glossaries/name/{name}/importRdf`	`EditAll` on the glossary	`format` (`turtle`, `rdfxml`, `ntriples`, `jsonld`), `dryRun` (default `true`).
`GET /api/v1/glossaries/{id}/exportOntology`	`ViewAll`	`format`, `includeRelations` (default `true`).
`POST /api/v1/glossaries/{id}/sparql`	`ViewAll`	Read-only SPARQL scoped to that glossary. No `SERVICE`, no other assets.

Applications

Application	Purpose
`RdfIndexApp`, RDF Knowledge Graph Indexing	Initial index and full rebuilds. Config: `entities`, `recreateIndex` (default `true`), `batchSize` (100), `producerThreads` (2), `consumerThreads` (3), `queueSize` (5000), `useDistributedIndexing` (default `true`), `partitionSize` (10000).
`RdfInferenceApp`, RDF Inference Materialization	Materializes enabled inference rules into their durable per-rule named graphs. Runs on a schedule (every 5 minutes by default); use Run Now in Settings → Applications to materialize immediately.

# Trigger a full reindex
curl -X POST "$COLLATE_HOST/api/v1/apps/trigger/RdfIndexApp" -H "Authorization: Bearer $TOKEN"

# Status
curl "$COLLATE_HOST/api/v1/apps/name/RdfIndexApp/status" -H "Authorization: Bearer $TOKEN"

# Stop a long-running index
curl -X POST "$COLLATE_HOST/api/v1/apps/stop/RdfIndexApp" -H "Authorization: Bearer $TOKEN"

Only one reindex job can be active per cluster at a time.

Ontology Studio API

Ontology authoring endpoints.

MCP tools

The same capabilities, for agents.

Ontology & Knowledge Graph

Knowledge Graph API Reference | Collate

Knowledge Graph API Reference

Status

SPARQL

Entities and Graph

`GET /glossary/graph` Parameters

Validation

Inference Rules

Custom Ontology Extensions

Insights

Semantic Search

SQL → SPARQL

Linked Open Data

Glossary Ontology Endpoints

Applications

Next

Ontology Studio API

MCP tools

​Knowledge Graph API Reference

​Status

​SPARQL

​Entities and Graph

​GET /glossary/graph Parameters

​Validation

​Inference Rules

​Custom Ontology Extensions

​Insights

​Semantic Search

​SQL → SPARQL

​Linked Open Data

​Glossary Ontology Endpoints

​Applications

​Next

Ontology Studio API

MCP tools

Knowledge Graph API Reference

Status

SPARQL

Entities and Graph

`GET /glossary/graph` Parameters

Validation

Inference Rules

Custom Ontology Extensions

Insights

Semantic Search

SQL → SPARQL

Linked Open Data

Glossary Ontology Endpoints

Applications

Next