temporalio.contrib.openai_agents.StatelessMCPServerProvider

class documentation

class StatelessMCPServerProvider: (source)

Constructor: StatelessMCPServerProvider(name, server_factory)

A stateless MCP server implementation for Temporal workflows.

This class wraps a function to create MCP servers to make them stateless by executing each MCP operation as a separate Temporal activity. Each operation (list_tools, call_tool, etc.) will connect to the underlying server, execute the operation, and then clean up the connection.

This approach will not maintain state across calls. If the desired MCPServer needs persistent state in order to function, this cannot be used.

Method	`__init__`	Initialize the stateless temporal MCP server.
Property	`name`	Get the server name.
Method	`_create_server`	Undocumented
Method	`_get_activities`	Undocumented
Instance Variable	`_name`	Undocumented
Instance Variable	`_server_accepts_arguments`	Undocumented
Instance Variable	`_server_factory`	Undocumented

def __init__(self, name: str, server_factory: Callable[[], MCPServer] | Callable[[(Any | None)], MCPServer]): (source) ¶

Initialize the stateless temporal MCP server.

Parameters
name:`str`	The name of the MCP server.
server_factory:`Callable[[], MCPServer] \| Callable[[(Any \| None)], MCPServer]`	A function which will produce MCPServer instances. It should return a new server each time so that state is not shared between workflow runs.

@property
name: str = (source) ¶

Get the server name.

def _create_server(self, factory_argument: Any | None) -> MCPServer: (source) ¶

Undocumented

def _get_activities(self) -> Sequence[Callable]: (source) ¶

Undocumented

_name = (source) ¶

Undocumented

_server_accepts_arguments = (source) ¶

Undocumented

_server_factory = (source) ¶

Undocumented