AgentOptimization

Properties

Name	Type	Description	Notes
access	AiConfigsAccess		[optional]
links	ParentAndSelfLinks		[optional]
id	String	The agent optimization ID
key	String	The agent optimization key
aiConfigKey	String	The key of the AI Config to optimize
maxAttempts	Integer	The maximum number of attempts
modelChoices	List<String>	The models to try
judgeModel	String	The model used to judge responses
variableChoices	List<Map<String, Object>>	Sets of prompt variable values to try
acceptanceStatements	List<AgentOptimizationAcceptanceStatement>	Statements a response must satisfy to pass
judges	List<AgentOptimizationJudge>	Judges that score each response
userInputOptions	List<String>	User inputs to test with
groundTruthResponses	List<String>	Expected responses to compare against
metricKey	String	The key of the metric to optimize	[optional]
tokenLimit	Integer	The maximum number of tokens to use	[optional]
variationKey	String	The key of the variation to start from	[optional]
label	String	A label for the optimization	[optional]
latencyOptimization	Boolean	Whether to optimize for lower latency	[optional]
tokenOptimization	Boolean	Whether to optimize for lower token use	[optional]
autoCommit	Boolean	Whether to save the best result as a new variation automatically	[optional]
version	Integer	The version number of this resource
createdAt	Long	When this resource was created, in Unix milliseconds