Skip to content
Cloudflare API
Start here
API Reference
Radar
AI

Inference

Get Workers AI inference distribution by dimension
radar.ai.inference.summary_v2(Literal["MODEL", "TASK"]dimension, InferenceSummaryV2Params**kwargs) -> InferenceSummaryV2Response
GET/radar/ai/inference/summary/{dimension}
Get time series distribution of Workers AI inference by dimension.
radar.ai.inference.timeseries_groups_v2(Literal["MODEL", "TASK"]dimension, InferenceTimeseriesGroupsV2Params**kwargs) -> InferenceTimeseriesGroupsV2Response
GET/radar/ai/inference/timeseries_groups/{dimension}
ModelsExpand Collapse
class InferenceSummaryV2Response: …
meta: Meta

Metadata for the results.

confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
summary_0: Dict[str, str]
class InferenceTimeseriesGroupsV2Response: …
meta: Meta

Metadata for the results.

agg_interval: Literal["FIFTEEN_MINUTES", "ONE_HOUR", "ONE_DAY", 2 more]

Aggregation interval of the results (e.g., in 15 minutes or 1 hour intervals). Refer to Aggregation intervals.

One of the following:
"FIFTEEN_MINUTES"
"ONE_HOUR"
"ONE_DAY"
"ONE_WEEK"
"ONE_MONTH"
confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
serie_0: Serie0
timestamps: List[datetime]

InferenceSummary

Get Workers AI models summary
Deprecated
radar.ai.inference.summary.model(SummaryModelParams**kwargs) -> SummaryModelResponse
GET/radar/ai/inference/summary/model
Get Workers AI tasks summary
Deprecated
radar.ai.inference.summary.task(SummaryTaskParams**kwargs) -> SummaryTaskResponse
GET/radar/ai/inference/summary/task
ModelsExpand Collapse
class SummaryModelResponse: …
meta: Meta

Metadata for the results.

confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
summary_0: Dict[str, str]
class SummaryTaskResponse: …
meta: Meta

Metadata for the results.

confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
summary_0: Dict[str, str]

InferenceTimeseries Groups

InferenceTimeseries GroupsSummary

Get Workers AI models time series
Deprecated
radar.ai.inference.timeseries_groups.summary.model(SummaryModelParams**kwargs) -> SummaryModelResponse
GET/radar/ai/inference/timeseries_groups/model
Get Workers AI tasks time series
Deprecated
radar.ai.inference.timeseries_groups.summary.task(SummaryTaskParams**kwargs) -> SummaryTaskResponse
GET/radar/ai/inference/timeseries_groups/task
ModelsExpand Collapse
class SummaryModelResponse: …
meta: Meta

Metadata for the results.

agg_interval: Literal["FIFTEEN_MINUTES", "ONE_HOUR", "ONE_DAY", 2 more]

Aggregation interval of the results (e.g., in 15 minutes or 1 hour intervals). Refer to Aggregation intervals.

One of the following:
"FIFTEEN_MINUTES"
"ONE_HOUR"
"ONE_DAY"
"ONE_WEEK"
"ONE_MONTH"
confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
serie_0: Serie0
timestamps: List[datetime]
class SummaryTaskResponse: …
meta: Meta

Metadata for the results.

agg_interval: Literal["FIFTEEN_MINUTES", "ONE_HOUR", "ONE_DAY", 2 more]

Aggregation interval of the results (e.g., in 15 minutes or 1 hour intervals). Refer to Aggregation intervals.

One of the following:
"FIFTEEN_MINUTES"
"ONE_HOUR"
"ONE_DAY"
"ONE_WEEK"
"ONE_MONTH"
confidence_info: MetaConfidenceInfo
annotations: List[MetaConfidenceInfoAnnotation]
data_source: Literal["ALL", "AI_BOTS", "AI_GATEWAY", 22 more]

Data source for annotations.

One of the following:
"ALL"
"AI_BOTS"
"AI_GATEWAY"
"BGP"
"BOTS"
"CONNECTION_ANOMALY"
"CT"
"DNS"
"DNS_MAGNITUDE"
"DNS_AS112"
"DOS"
"EMAIL_ROUTING"
"EMAIL_SECURITY"
"FW"
"FW_PG"
"HTTP"
"HTTP_CONTROL"
"HTTP_CRAWLER_REFERER"
"HTTP_ORIGINS"
"IQI"
"LEAKED_CREDENTIALS"
"NET"
"ROBOTS_TXT"
"SPEED"
"WORKERS_AI"
description: str
end_date: datetime
formatdate-time
event_type: Literal["EVENT", "GENERAL", "OUTAGE", 3 more]

Event type for annotations.

One of the following:
"EVENT"
"GENERAL"
"OUTAGE"
"PARTIAL_PROJECTION"
"PIPELINE"
"TRAFFIC_ANOMALY"
is_instantaneous: bool

Whether event is a single point in time or a time range.

linked_url: str
formaturi
start_date: datetime
formatdate-time
level: int

Provides an indication of how much confidence Cloudflare has in the data.

date_range: List[MetaDateRange]
end_time: datetime

Adjusted end of date range.

formatdate-time
start_time: datetime

Adjusted start of date range.

formatdate-time
last_updated: datetime

Timestamp of the last dataset update.

formatdate-time
normalization: Literal["PERCENTAGE", "MIN0_MAX", "MIN_MAX", 5 more]

Normalization method applied to the results. Refer to Normalization methods.

One of the following:
"PERCENTAGE"
"MIN0_MAX"
"MIN_MAX"
"RAW_VALUES"
"PERCENTAGE_CHANGE"
"ROLLING_AVERAGE"
"OVERLAPPED_PERCENTAGE"
"RATIO"
units: List[MetaUnit]

Measurement units for the results.

name: str
value: str
serie_0: Serie0
timestamps: List[datetime]