LM-Kit OneDocs2026.9.7lm-kit.comEULA
API Compatibility

OpenAI · Vector Stores

Lists the caller's vector stores.#

GET/v1/vector_stores

Responses

StatusTypeDescription
200

OK

curl -X GET "$LMKIT_ONE_URL/v1/vector_stores" \
  -H "Authorization: Bearer $LMKIT_API_KEY"

Creates a vector store.#

POST/v1/vector_stores

Request body

text/json ·

PropertyTypeDescription
namestring

The vector store's display name.

Responses

StatusTypeDescription
200

OK

curl -X POST "$LMKIT_ONE_URL/v1/vector_stores" \
  -H "Authorization: Bearer $LMKIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "string"
}'

Retrieves a vector store.#

GET/v1/vector_stores/{storeId}

Parameters

NameInTypeDescription
storeIdrequiredpathstring

Responses

StatusTypeDescription
200

OK

curl -X GET "$LMKIT_ONE_URL/v1/vector_stores/$STOREID" \
  -H "Authorization: Bearer $LMKIT_API_KEY"

Deletes a vector store and everything indexed in it.#

DELETE/v1/vector_stores/{storeId}

Parameters

NameInTypeDescription
storeIdrequiredpathstring

Responses

StatusTypeDescription
200

OK

curl -X DELETE "$LMKIT_ONE_URL/v1/vector_stores/$STOREID" \
  -H "Authorization: Bearer $LMKIT_API_KEY"

Lists the files attached to a vector store.#

GET/v1/vector_stores/{storeId}/files

Newest first (order=asc for oldest first). 'filter' keeps one status ('in_progress' or 'completed'), 'after' and 'before' resume around a file id, 'limit' caps the page (1 to 100, default 20), and has_more says whether a further page exists.

Parameters

NameInTypeDescription
storeIdrequiredpathstring
limitqueryobject (int32)

Page size, 1 to 100.

orderquerystring

'asc' or 'desc' (default) by indexing time.

afterquerystring

The file id to resume after.

beforequerystring

The file id to stop before.

filterquerystring

Only files in this status.

Responses

StatusTypeDescription
200

OK

curl -X GET "$LMKIT_ONE_URL/v1/vector_stores/$STOREID/files" \
  -H "Authorization: Bearer $LMKIT_API_KEY"

Attaches an uploaded file to a vector store, indexing it for retrieval.#

POST/v1/vector_stores/{storeId}/files

The file (from POST /v1/files) is parsed, chunked, and indexed by the managed ingestion pipeline. A file whose status is still 'in_progress' is already full-text searchable and reports 'completed' once its embedding pass has landed; poll GET .../files/ the way OpenAI SDK helpers do.

Parameters

NameInTypeDescription
storeIdrequiredpathstring

Request body

text/json ·

PropertyTypeDescription
file_idstring

The id of an uploaded file (from POST /v1/files) to index into this store.

Responses

StatusTypeDescription
200

OK

curl -X POST "$LMKIT_ONE_URL/v1/vector_stores/$STOREID/files" \
  -H "Authorization: Bearer $LMKIT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "file_id": "string"
}'

Retrieves a file attached to a vector store, with its indexing status.#

GET/v1/vector_stores/{storeId}/files/{fileId}

Parameters

NameInTypeDescription
storeIdrequiredpathstring
fileIdrequiredpathstring

Responses

StatusTypeDescription
200

OK

curl -X GET "$LMKIT_ONE_URL/v1/vector_stores/$STOREID/files/$FILEID" \
  -H "Authorization: Bearer $LMKIT_API_KEY"

Detaches a file from a vector store, removing what was indexed from it.#

DELETE/v1/vector_stores/{storeId}/files/{fileId}

The uploaded file itself stays in the caller's bucket; delete it through DELETE /v1/files/.

Parameters

NameInTypeDescription
storeIdrequiredpathstring
fileIdrequiredpathstring

Responses

StatusTypeDescription
200

OK

curl -X DELETE "$LMKIT_ONE_URL/v1/vector_stores/$STOREID/files/$FILEID" \
  -H "Authorization: Bearer $LMKIT_API_KEY"