Public research interface · Version 1.0.0

Knowledge Graph API

Read-only access to the reader’s V3 claims, thematic categories, and related entity or relation labels. Responses retain source-page provenance.

Base URL/api/v1
Page sizeUp to 100 records

Endpoints

All endpoints are read-only.

GET /api/v1/tags

Find raw entity or relation labels before filtering claims.

GET /api/v1/tags/{kind}/{tag_id}/similar

Rank same-kind labels by embedding similarity.

GET /api/v1/triples

Filter provenance-preserving V3 claims.

GET /api/v1/triples/{claim_id}

Retrieve one claim with its source context.

GET /api/v1/thematic-patterns

Aggregate matching claims by thematic category pattern.

GET /api/v1/categories

List the controlled entity and relation categories.

Query rules

Raw labels

Use subject, predicate, object, or the corresponding *_contains filter. Repeat a filter to form an OR union.

Categories

Use subject_category, relation_category, and object_category for thematic patterns. Category values combine with other filter types.

Source scope

Limit claims to volume and page, an inclusive start_page/end_page range, or one sentence_id.

Paging

Pass pagination.nextCursor back as cursor with unchanged filters. Use include_text=true only when sentence text is needed.

Examples

# Find a raw label
curl "https://burmeseneuralreader.com/api/v1/tags?kind=entity&q=Hsinbyushin&limit=20"

# Find claims for an entity tag
curl "https://burmeseneuralreader.com/api/v1/triples?entity_tag_id=ENTITY_TAG_ID&limit=25"

# Aggregate a page range by thematic pattern
curl "https://burmeseneuralreader.com/api/v1/thematic-patterns?volume=vol2&start_page=235&end_page=245&limit=25"
Scope Collection pages contain at most 100 records. Queries are bounded and read-only; arbitrary SPARQL is not exposed.