Usage

Authorizations

Authorization

string

header

required

When using Friendli Suite API, you need to provide a Personal API Key (e.g. flp_XXX) as a Bearer token for authentication and authorization purposes.

For more detailed information, please refer here.

Headers

X-Friendli-Team

string

ID of team to run requests as (optional parameter).

Query Parameters

start_time

string<date-time>

required

RFC 3339 timestamp in UTC. Must be within the past 365 days and align with the selected bucket_width (e.g., 11:05:00Z for 5m, 11:00:00Z for 1h).

end_time

string<date-time>

required

RFC 3339 timestamp in UTC. Must be later than start_time and align with the selected bucket_width (e.g., 11:10:00Z for 5m, 12:00:00Z for 1h).

bucket_width

enum<string>

default:1d

Width of each time bucket. Must be selected based on the query duration (end_time - start_time) and must not exceed the query duration.

Available options:

5m,

15m,

30m,

45m,

1h,

6h,

12h,

1d

limit

integer

Specifies the number of buckets to return.

bucket_width=1d: default 7, max 31
bucket_width=12h: default 2, max 14
bucket_width=6h: default 4, max 28
bucket_width=1h: default 24, max 168
bucket_width=45m: default 2, max 32
bucket_width=30m: default 2, max 48
bucket_width=15m: default 4, max 96
bucket_width=5m: default 12, max 288

page

string

A cursor for use in pagination. Corresponding to the next_page field from the previous response.

group_by

enum<string>[]

Group the usage by the specified fields. Support fields include model, product_type, user_id and any combination of them.

Usage group-by option.

Available options:

model,

product_type,

user_id

models

string[]

Return only usage for the listed Model APIs and Dedicated Endpoints. Values are model IDs, endpoint IDs, or a mix.

user_ids

string[]

Return only usage for these users.

product_types

enum<string>[]

Return only usage for the specified product types. Supported fields include model_apis, dedicated_endpoints and any combination of them.

Product type.

Available options:

model_apis,

dedicated_endpoints

gpu_types

string[]

Return only usage for the specified GPU types. When this filter is given, gpu_usage and reserved_gpu_usage will be provided if present; request, token, and audio usage fields are omitted. Combining gpu_types with product_types=model_apis is rejected with HTTP 400.

capacity_types

enum<string>[]

Return only usage for the specified capacity types. Supported fields include reservation only. When this filter is given, only reserved_gpu_usage will be provided; gpu_usage, request, token, and audio usage fields are omitted. Combining capacity_types with product_types=model_apis is rejected with HTTP 400.

GPU capacity type.

Available options:

reservation

Response

Team usage response

Usage response.

data

UsageBucket · object[]

required

The items in this page.

Hide child attributes

data.start_time

string<date-time>

required

data.end_time

string<date-time>

required

data.results

UsageResult · object[]

required

Hide child attributes

data.results.product_type

enum<string> | null

When group_by=product_type, this field provides the product type of the grouped usage result. Supported values are model_apis and dedicated_endpoints.

Available options:

model_apis,

dedicated_endpoints

Example:

"model_apis"

data.results.model

string | null

When group_by=model, this field provides the model ID of the grouped usage result. For Model APIs usage, this is the model ID. For Dedicated Endpoints usage, this is the endpoint ID.

Example:

"zai-org/GLM-5.3"

data.results.user_id

string | null

When group_by=user_id, this field provides the user ID of the grouped usage result.

Example:

"AvBI1gXTuRux"

data.results.num_model_requests

integer | null

default:0

The count of requests made.

Example:

1200

data.results.input_tokens

integer | null

default:0

The aggregated number of input tokens used, including cached tokens.

Example:

180000

data.results.input_cached_tokens

integer | null

default:0

The aggregated number of input tokens that has been cached.

Example:

45000

data.results.output_tokens

integer | null

default:0

The aggregated number of text output tokens used.

Example:

62000

data.results.gpu_usage

Gpu Usage · object | null

Total GPU time in seconds by GPU type, across both on-demand and reserved capacity. Missing keys mean zero. With group_by=user_id, usage is attributed to the user who created the endpoint.

Hide child attributes

data.results.gpu_usage.{key}

integer

Example:

{ "b200": 606, "b300": 844 }

data.results.reserved_gpu_usage

Reserved Gpu Usage · object | null

Reserved GPU time in seconds by GPU type, covered by your reservation during the reservation period. Included in gpu_usage; do not add them together. Missing keys mean zero. Usage is reported per endpoint, but which endpoints consume reserved GPU time is determined automatically and cannot be configured. With group_by=user_id, usage is attributed to the user who created the endpoint.

Hide child attributes

data.results.reserved_gpu_usage.{key}

integer

Example:

{ "b200": 400 }

data.results.processed_audio_length_ms

integer | null

default:0

The number of processed audio length in milliseconds. Zero when there is no audio usage.

Example:

300000

has_more

boolean

required

Whether more items are available beyond this page.

Example:

false

next_page

string | null

A cursor to fetch the next page. Null when there are no more results.

API Reference

Model APIs

Dedicated

Administration

Usage

Authorizations

Headers

Query Parameters

Response