Get Dataset Memory Status
Get the memory state of datasets: what each one can answer right now.
cognify() (and remember()) build memory in two runs per dataset: indexing (chunk search) and cognify (the knowledge graph). This endpoint folds their latest runs into one state per dataset, so a client can tell whether a search will answer and how good the answers are, instead of reading two pipeline statuses.
Query Parameters
- dataset (List[UUID]): Dataset UUIDs to check (from GET /api/v1/datasets). Omit to get the state of all datasets you can read.
Response
{dataset_id: {state, search_level, searchable, graph_ready, message, indexing_status, cognify_status, error_class, error_message}}. Every requested
dataset is present; state is empty when nothing ran on it.
- state:
empty,staged,indexing,indexed,indexing_failed,extracting,readyorextraction_failed. During a cognify() build it goesindexing,extracting, thenready. - search_level:
none,chunks(CHUNKS, CHUNKS_LEXICAL, RAG_COMPLETION, HYBRID_COMPLETION answer) orgraph(everything answers) - message: one sentence saying what answers now and what improves next
Error Codes
- 403 Forbidden: The request owner cannot read every requested dataset
- 409 Conflict: An unexpected error occurred while deriving the state
Authorizations
Bearer authentication header of the form Bearer <token>, where <token> is your auth token.
Query Parameters
Dataset UUIDs to check (from GET /api/v1/datasets). Omit to get status for all datasets you can read.
Response
Successful Response
A dataset's memory state, folded from its indexing and cognify runs.