nat.plugins.mcp.client.client_config

Classes

MCPToolOverrideConfig	Configuration for overriding tool properties when exposing from MCP server.
MCPServerConfig	Server connection details for MCP client.
MCPClientBaseConfig	Base configuration shared by MCP client variants.
MCPClientConfig	Configuration for connecting to an MCP server as a client and exposing selected tools.
PerUserMCPClientConfig	MCP Client configuration for per-user workflows that are registered with @register_per_user_function,

Module Contents

class MCPToolOverrideConfig(/, **data: Any)

Bases: pydantic.BaseModel

Configuration for overriding tool properties when exposing from MCP server.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

alias: str | None = None

description: str | None = None

class MCPServerConfig(/, **data: Any)

Bases: pydantic.BaseModel

Server connection details for MCP client. Supports stdio, sse, and streamable-http transports. streamable-http is the recommended default for HTTP-based connections.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

transport: Literal['stdio', 'sse', 'streamable-http'] = None

url: pydantic.HttpUrl | None = None

command: str | None = None

args: list[str] | None = None

env: dict[str, str] | None = None

auth_provider: str | nat.data_models.component_ref.AuthenticationRef | None = None

validate_model()

Validate that stdio and SSE/Streamable HTTP properties are mutually exclusive.

class MCPClientBaseConfig(/, **data: Any)

Bases: nat.data_models.function.FunctionGroupBaseConfig

Base configuration shared by MCP client variants.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

server: MCPServerConfig = None

tool_call_timeout: datetime.timedelta = None

auth_flow_timeout: datetime.timedelta = None

reconnect_enabled: bool = None

reconnect_max_attempts: int = None

reconnect_initial_backoff: float = None

reconnect_max_backoff: float = None

tool_overrides: dict[str, MCPToolOverrideConfig] | None = None

_validate_reconnect_backoff() → MCPClientBaseConfig

Validate reconnect backoff values.

class MCPClientConfig(/, **data: Any)

Bases: MCPClientBaseConfig

Configuration for connecting to an MCP server as a client and exposing selected tools.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

session_aware_tools: bool = None

max_sessions: int = None

session_idle_timeout: datetime.timedelta = None

class PerUserMCPClientConfig(/, **data: Any)

Bases: MCPClientBaseConfig

MCP Client configuration for per-user workflows that are registered with @register_per_user_function,

and each user gets their own MCP client instance.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.