InferenceDetails interface
Details for an inference call
Properties
| cache |
Number of input tokens read from a cache |
| cache |
Number of input tokens written to a cache |
| endpoint | The endpoint for the inference call |
| finish |
Array of finish reasons |
| input |
Number of input tokens used |
| model | The requested model name/identifier. At least one of model or responseModel is required. |
| operation |
The operation name/type for the inference |
| output |
Number of output tokens generated |
| provider |
The provider name (e.g., openai, azure, anthropic) |
| response |
The model name/identifier returned by the provider |
| thought |
The thought process used by the agent |
Property Details
cacheReadInputTokens
Number of input tokens read from a cache
cacheReadInputTokens?: numberProperty Value
number
cacheWriteInputTokens
Number of input tokens written to a cache
cacheWriteInputTokens?: numberProperty Value
number
endpoint
finishReasons
Array of finish reasons
finishReasons?: string[]Property Value
string[]
inputTokens
Number of input tokens used
inputTokens?: numberProperty Value
number
model
The requested model name/identifier. At least one of model or responseModel is required.
model?: stringProperty Value
string
operationName
The operation name/type for the inference
operationName: InferenceOperationTypeProperty Value
outputTokens
Number of output tokens generated
outputTokens?: numberProperty Value
number
providerName
The provider name (e.g., openai, azure, anthropic)
providerName: stringProperty Value
string
responseModel
The model name/identifier returned by the provider
responseModel?: stringProperty Value
string
thoughtProcess
The thought process used by the agent
thoughtProcess?: stringProperty Value
string