List Usage Records

curl -X GET 'https://app.memorylake.ai/openapi/memorylake/admin/v1/usage/records' \
  -H 'Authorization: Bearer sk_xxxxxx'

{
  "success": true,
  "message": "Operation completed successfully",
  "data": {
    "items": [
      {
        "completion_tokens": 1200000,
        "created_at": 1760000000,
        "id": "a672d31430497d2bf7e0a35c08f0d757",
        "kind": "model_call",
        "model": "example",
        "prompt_tokens": 1200000,
        "used_tokens": 1200000,
        "member": {
          "principal_id": "d20a9584eebe672f5820bb4ad5859745",
          "display_name": "Alice"
        },
        "request_id": "c5ed51a925f042a4ae5fa015330a6345",
        "trace_id": "95917fd92abee21729b9575f0b9c9cec"
      }
    ],
    "continuation_token": "eyJvZmZzZXQiOjIwfQ",
    "total": 1
  }
}

curl -X GET 'https://app.memorylake.ai/openapi/memorylake/admin/v1/usage/records' \
  -H 'Authorization: Bearer sk_xxxxxx'

{
  "success": true,
  "message": "Operation completed successfully",
  "data": {
    "items": [
      {
        "completion_tokens": 1200000,
        "created_at": 1760000000,
        "id": "a672d31430497d2bf7e0a35c08f0d757",
        "kind": "model_call",
        "model": "example",
        "prompt_tokens": 1200000,
        "used_tokens": 1200000,
        "member": {
          "principal_id": "d20a9584eebe672f5820bb4ad5859745",
          "display_name": "Alice"
        },
        "request_id": "c5ed51a925f042a4ae5fa015330a6345",
        "trace_id": "95917fd92abee21729b9575f0b9c9cec"
      }
    ],
    "continuation_token": "eyJvZmZzZXQiOjIwfQ",
    "total": 1
  }
}

GET /openapi/memorylake/admin/v1/usage/records

Lists individual consumption records, newest first: one per billed model request or platform operation, with tokens and the consumption charged in tokens. Consumption is attributed to members — to tell applications apart, give each its own virtual member. Records are near real-time; Get Usage reads aggregates refreshed every few minutes, so the two can briefly disagree. With log:read_all the list covers the team; a caller holding only log:read_self sees their own records, and filtering by another member’s principal_id is refused. The window defaults to the last 7 days and may span at most 92 days. Unlike other lists this one returns no total — page with continuation_token until it is absent. Rate limit: 20 requests per minute per key by default, counted separately from all other operations. When paging through many records, use the largest page_size (100).

Permission required: log:read_all (whole team) or log:read_self (your own only)

When you filter by another member’s principal_id: log:read_all

Query Parameters

integer

default:"20"

Number of items per page (1–100)

string

Token from the previous page. Opaque — do not construct or parse it. Send the same filters with it.

integer

Earliest record time to include, Unix seconds (inclusive). Defaults to 7 days before end_time.

integer

Latest record time, Unix seconds (exclusive). Defaults to now. The window may span at most 92 days.

array

Only records for these models or platform operations (files_process, memory_input, retrieval_call, search_call). Repeat the parameter for several.

string

model_call: model requests. platform_task: platform operations such as file processing and memory writes. Omit for both.

array

Only records of these members. Repeat the parameter for several. A caller limited to their own records may only pass their own id.

string

Only the record of this request id.

curl -X GET 'https://app.memorylake.ai/openapi/memorylake/admin/v1/usage/records' \
  -H 'Authorization: Bearer sk_xxxxxx'

Response

object

Show Load More Response Usage Record object

array

List of items in the current page

Show Usage Record object

integer

Completion (output) tokens

integer

When the consumption was recorded, Unix seconds

string

Record id

string

model_call or platform_task

string

Model name, or the platform operation (files_process, memory_input, retrieval_call, search_call)

integer

Prompt (input) tokens

integer

Consumption charged against the balance, in tokens

object

The member the consumption is attributed to; absent on records that carry none

Show Usage Record Member object

string

The member’s principal id

string

The member’s display name

string

Request id, when one was recorded

string

Trace id, when one was recorded

string

Token to fetch the next page; absent on the last page

integer

Total number of items matching the filter across all pages; omitted when the count is unavailable

{
  "success": true,
  "message": "Operation completed successfully",
  "data": {
    "items": [
      {
        "completion_tokens": 1200000,
        "created_at": 1760000000,
        "id": "a672d31430497d2bf7e0a35c08f0d757",
        "kind": "model_call",
        "model": "example",
        "prompt_tokens": 1200000,
        "used_tokens": 1200000,
        "member": {
          "principal_id": "d20a9584eebe672f5820bb4ad5859745",
          "display_name": "Alice"
        },
        "request_id": "c5ed51a925f042a4ae5fa015330a6345",
        "trace_id": "95917fd92abee21729b9575f0b9c9cec"
      }
    ],
    "continuation_token": "eyJvZmZzZXQiOjIwfQ",
    "total": 1
  }
}

Errors

Errors answer with success: false, a message and an error_code. See Error Codes for every code.

Status	Meaning
`400`	A parameter or the body is invalid
`401`	The API key is missing, invalid, disabled or expired
`403`	The key is not allowed to do this
`422`	The body or a parameter failed validation
`429`	Too many requests. Wait for the time in `Retry-After`
`500`	Server-side error

MemoryLake API

Memory Router API

Query Parameters

Response

Errors

​Query Parameters

​Response

​Errors

Query Parameters

Response

Errors