Update entity
Find duplicate entity candidates
Returns groups of entities that are potential duplicates based on name similarity. Uses fuzzy string matching to identify entities that likely refer to the same thing but were not caught by automatic resolution during extraction. One scan runs at a time (single-flight); identical requests join the running scan. On large graphs the scan may outlast the inline wait window, in which case the response is 202 with a progress fraction — poll the same URL (without refresh) until it returns status "complete". Completed results are cached server-side; entity merges invalidate the cache.
query Parameters
thresholdSimilarity threshold for duplicate detection (0.5 to 1.0)
limitMaximum number of duplicate groups to return
refreshBypass the cached result and force a fresh scan
Find duplicate entity candidates › Responses
Duplicate candidates retrieved (status "complete")
statusAlways "complete" on a 200 response
cachedTrue when served from the server-side scan cache
total_groupsTotal number of duplicate groups found
thresholdSimilarity threshold used for detection
Merge duplicate entities
Merges one or more duplicate entities into a canonical entity. All relationships, chunk mentions, and community memberships from the merged entities are transferred to the canonical entity. Merged entities are removed from the graph.
Merge duplicate entities › Request Body
canonicalThe name of the entity to keep (canonical/primary entity)
mergeList of entity names to merge into the canonical entity
Merge duplicate entities › Responses
Entities merged successfully
canonicalThe canonical entity name that was kept
mergedEntity names that were merged and removed
relationships_transferredNumber of relationships moved to the canonical entity
mentions_transferredNumber of chunk mentions moved to the canonical entity
aliases_addedAliases added to the canonical entity from merged entities
Get entity merge history
Returns a chronological log of entity merge operations, including which entities were merged into which canonical entity and when.
query Parameters
limitMaximum number of history entries to return
Get entity merge history › Responses
Merge history retrieved
totalTotal number of merge history entries
List relationships
List relationships between entities with server-side pagination and search. Excludes internal edge types (MENTIONS, HAS_MEMBER, FROM_DOCUMENT, CO_MENTION). For restricted API keys, results are scoped to accessible collections.
query Parameters
rel_typeFilter by relationship type
searchSearch in source, target, and description
skiplimitList relationships › Responses
Relationships retrieved
totalTotal number of matching relationships
Analyze cross-document relations
Triggers Phase B of the extraction pipeline. The main (large) model analyzes all entities and discovers cross-document relations. Run this after batch document processing to build the relationship graph.
query Parameters
collection_idScope to a specific collection (None = global)
scope'recent' for new entities, 'full' for all
Analyze cross-document relations › Responses
Analysis task started
task_idstatusmessagetipClean up orphaned entities and communities
Remove entities not connected to any document chunks and communities with no member entities. Use this after bulk document deletion to clean up stale graph data.
Clean up orphaned entities and communities › Responses
Cleanup completed
messageorphaned_entities_removedorphaned_communities_removed