OutputToolBinding Class — langchain Architecture

Source Code

libs/langchain_v1/langchain/agents/structured_output.py lines 308–359

class OutputToolBinding(Generic[SchemaT]):
    """Information for tracking structured output tool metadata.

    This contains all necessary information to handle structured responses generated via
    tool calls, including the original schema, its type classification, and the
    corresponding tool implementation used by the tools strategy.
    """

    schema: type[SchemaT] | dict[str, Any]
    """The original schema provided for structured output (Pydantic model, dataclass,
    TypedDict, or JSON schema dict).
    """

    schema_kind: SchemaKind
    """Classification of the schema type for proper response construction."""

    tool: BaseTool
    """LangChain tool instance created from the schema for model binding."""

    @classmethod
    def from_schema_spec(cls, schema_spec: _SchemaSpec[SchemaT]) -> Self:
        """Create an `OutputToolBinding` instance from a `SchemaSpec`.

        Args:
            schema_spec: The `SchemaSpec` to convert

        Returns:
            An `OutputToolBinding` instance with the appropriate tool created
        """
        return cls(
            schema=schema_spec.schema,
            schema_kind=schema_spec.schema_kind,
            tool=StructuredTool(
                args_schema=schema_spec.json_schema,
                name=schema_spec.name,
                description=schema_spec.description,
            ),
        )

    def parse(self, tool_args: dict[str, Any]) -> SchemaT:
        """Parse tool arguments according to the schema.

        Args:
            tool_args: The arguments from the tool call

        Returns:
            The parsed response according to the schema type

        Raises:
            ValueError: If parsing fails
        """
        return _parse_with_schema(self.schema, self.schema_kind, tool_args)