Mockers

relai.mocker.Persona(user_persona, intent=None, starting_message=None, trajectory=None, identifier_in_trajectory='user', tools=None, model='gpt-5-mini', reasoning_effort=None, extra_model_args=None)

Bases: BaseMocker

A mocker class for creating AI personas with specific behaviors and tools. A persona can be used to mimic a particular role by defining a system prompt and optionally equipping it with tools.

Attributes:

Name	Type	Description
user_persona	str	The description of the persona's characteristics and behavior.
intent	str | None	The intent or goal of the persona during interactions.
starting_message	str | None	An optional initial message that the persona will use to start interactions.
tools	Optional[list[Tool]]	A list of tools that the persona can use.

Initializes the Persona with a description, intent, optional starting message, tools, and model.

Parameters:

Name	Type	Description	Default
user_persona	str	The description of the persona's characteristics and behavior.	required
intent	str | None	The intent or goal of the persona during interactions.	None
starting_message	str | None	An optional initial message that the persona will use to start interactions.	None
trajectory	str | None	If provided, the persona will try to mimic the behavior of a certain entity (defaults to "user") recorded in the trajectory. The entity to mimic can be configured through identifier_in_trajectory.	None
identifier_in_trajectory	str	A string to identify the entity to mimic when a trajectory is provided. Defaults to "user".	'user'
tools	Optional[list[Tool]]	A list of tools that the persona can use.	None
model	str | LitellmModel	The AI model to use for simulating the persona's behavior. This can be a string identifier for OpenAI models (e.g. gpt-5-mini) or a LitellmModel (from agents.extensions.models.litellm_model import LitellmModel). For a full list of models supported in LiteLLM, see https://docs.litellm.ai/docs/providers	'gpt-5-mini'
reasoning_effort	Literal['none', 'minimal', 'low', 'medium', 'high', 'xhigh', 'default'] | None	The level of reasoning effort to use for the LLM, if supported by the provider.	None
extra_model_args	dict[str, Any] | None	Arbitrary keyword arguments to pass directly to the underlying model's API. Note that not all models support all parameters.	None

relai.mocker.MockTool(model='gpt-5-mini', reasoning_effort=None, extra_model_args=None, context=None)

Bases: BaseMocker

A mocker class for simulating the behavior of a tool used by an AI agent.

Initializes the MockTool with an optional model specification.

Parameters:

Name	Type	Description	Default
model	str | LitellmModel	The AI model to use for simulating the tool's behavior. This can be a string identifier for OpenAI models (e.g. gpt-5-mini) or a LitellmModel (from agents.extensions.models.litellm_model import LitellmModel). For a full list of models supported in LiteLLM, see https://docs.litellm.ai/docs/providers	'gpt-5-mini'
reasoning_effort	Literal['none', 'minimal', 'low', 'medium', 'high', 'xhigh', 'default'] | None	The level of reasoning effort to use for the LLM, if supported by the provider.	None
extra_model_args	dict[str, Any] | None	Arbitrary keyword arguments to pass directly to the underlying model's API. Note that not all models support all parameters.	None
context	str | None	Additional context to guide behavior of the mock tool.	None