You can monitor what your agents are doing in real time.
Access the Dashboard: Open your web browser and navigate to http://localhost:5006
You need to set up dashboard when creating agent: with dashboard=True parameter:
agent = Agent(
name=”MyAgent”,
instructions=”You are a helpful assistant.”,
functions=[],
dashboard=True # Enable the dashboard
)
Of course I assume you have your swarm buzzing happily around
There are all agent parameter below:
• name (Type: str): The name of the agent. Default is "Agent".
• model (Type: str): Specifies the language model to be used by the agent. Default is "gpt-4o".
• instructions (Type: Union[str, Callable[[], str]]): A string or callable that provides instructions for the agent’s behavior. Default is "You are a helpful agent.".
• functions (Type: ListAgentFunction): A list of functions that the agent can call to perform tasks.
• tool_choice (Type: str): Determines how the agent chooses tools:
• "required": Forces the LLM to choose one of the provided functions.
• "auto": Lets the LLM decide if any tool needs to be called.
• "none": No function calls will be made.
• parallel_tool_calls (Type: bool): If set to True, allows the agent to call multiple functions simultaneously. Default is True.
• context_variables (Type: dict): A dictionary of additional context variables available for functions and agent instructions. Default is {}.
• max_turns (Type: int): The maximum number of conversational turns allowed. Default is infinity (float("inf")).
• model_override (Type: str): An optional string to override the model being used by the agent. Default is None.
• execute_tools (Type: bool): If set to False, interrupts execution and immediately returns a message when an agent tries to call a function. Default is True.
• stream (Type: bool): If set to True, enables streaming responses from the agent. Default is False.
• debug (Type: bool): Enables debugging mode for additional insights during execution.
• id (Type: str): A unique identifier for the agent instance.
• llm (Type: LanguageModelInstance): The specific language model instance used by the agent.
• template (Type: str): A template used for formatting responses.
• max_loops (Type: int): Maximum number of loops the agent can run.
• stopping_condition (Type: Callable): A callable function that determines when to stop looping.
• loop_interval (Type: float): Interval in seconds between loops.
• retry_attempts (Type: int): Number of retry attempts for failed LLM calls.
• retry_interval (Type: float): Interval in seconds between retry attempts.
• return_history (Type: bool): Indicates whether to return conversation history.
• stopping_token (Type: str): A token that stops the agent from looping when present in the response.
• dynamic_loops (Type: bool): Allows dynamic determination of loop counts based on conditions.
• interactive (Type: bool): Indicates whether to run in interactive mode.
• dashboard (Type: bool): Indicates whether to display a dashboard during operation.
These parameters provide flexibility in defining how an Agent behaves and interacts with users or other agents
