Documentation Index Fetch the complete documentation index at: https://docs.memorylake.ai/llms.txt
Use this file to discover all available pages before exploring further.
GET /openapi/memorylake/api/v2/projects/{id}/memories/conflicts/{conflictId}
Returns the full record for one conflict. You always get the memory snapshots captured at detection time. For memory-vs-document (m2d) conflicts, you also get the document excerpts that contradict the memory — excerpts whose source document has since been removed from the project are dropped.
Path Parameters curl -X GET 'https://app.memorylake.ai/openapi/memorylake/api/v2/projects/proj-a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6/memories/conflicts/conf-abc123' \ -H 'Authorization: Bearer sk_xxxxxx'
Response Longer explanation of the contradiction
m2m (memory vs memory) or m2d (memory vs document)
What kind of inconsistency:
logical — contradiction or tension between statements.knowledge — disagreement about facts or what was cited. Whether the conflict has been resolved
Resolution record — present only when resolved is true. Strategy applied: keep_memory, trust_memory, trust_document, or dismiss
Memory kept when the strategy required a choice
Memories forgotten as part of resolution
Actor that resolved the conflict
When the conflict was resolved
Whether underlying content has changed since detection
Memories involved in the conflict
Snapshots of each memory’s text at the moment the conflict was detected History revision ID, if any
Memory text at detection time
Document excerpts — non-empty when category is m2d ID of the project document
Display name of the document
Matched text span from the document Optional location hint (e.g. page or cell range)
ID of the event that triggered detection
{ "success" : true , "data" : { "id" : "conf-abc123" , "name" : "Contradicting restaurant preferences" , "description" : "An earlier memory says the user prefers spicy food; a later one says they avoid it." , "category" : "m2m" , "conflict_type" : "logical" , "resolved" : false , "stale" : false , "memory_ids" : [ "mem-001" , "mem-002" ], "memory_snapshots" : [ { "memory_id" : "mem-001" , "memory_text" : "User prefers spicy food" }, { "memory_id" : "mem-002" , "memory_text" : "User avoids spicy food" } ], "file_chunks" : [], "event_id" : "evt-xyz" , "created_at" : "2024-01-15T10:30:00Z" , "updated_at" : "2024-01-15T10:30:00Z" } }
Once you have the conflict detail, apply a decision with Resolve Memory Conflict . The returned forgotten_memory_ids tells you which memories were removed so your UI can stay in sync. 
