pydantic/pydantic · error · PydanticUserError

type-adapter-config-unused

type-adapter-config-unused

Error message

Cannot use `config` when the type is a BaseModel, dataclass or TypedDict. These types can have their own config and setting the config via the `config` parameter to TypeAdapter will not override it, thus the `config` you passed to TypeAdapter becomes meaningless, which is probably not what you want.

What it means

TypeAdapter accepts a config parameter, but only for types that don't carry their own config. _type_has_config returns True for BaseModel, pydantic dataclass, and TypedDict subclasses, which define config internally. Passing config for such a type would be silently ignored, so pydantic raises PydanticUserError code 'type-adapter-config-unused' instead.

Source

Thrown at pydantic/type_adapter.py:203

    def __init__(
        self,
        type: Any,
        *,
        config: ConfigDict | None = ...,
        _parent_depth: int = ...,
        module: str | None = ...,
    ) -> None: ...

    def __init__(
        self,
        type: Any,
        *,
        config: ConfigDict | None = None,
        _parent_depth: int = 2,
        module: str | None = None,
    ) -> None:
        if _type_has_config(type) and config is not None:
            raise PydanticUserError(
                'Cannot use `config` when the type is a BaseModel, dataclass or TypedDict.'
                ' These types can have their own config and setting the config via the `config`'
                ' parameter to TypeAdapter will not override it, thus the `config` you passed to'
                ' TypeAdapter becomes meaningless, which is probably not what you want.',
                code='type-adapter-config-unused',
            )

        self._type = type
        self._config = config
        self._parent_depth = _parent_depth
        self.pydantic_complete = False

        parent_frame = self._fetch_parent_frame()
        if isinstance(type, types.FunctionType):
            # Special case functions, which are *not* pushed to the `NsResolver` stack and without this special case
            # would only have access to the parent namespace where the `TypeAdapter` was instantiated (if the function is defined
            # in another module, we need to look at that module's globals).
            if parent_frame is not None:

View on GitHub (pinned to 2e5f0e2b42)

Solutions

Set the config on the model itself: class MyModel(BaseModel): model_config = ConfigDict(strict=True).
For plain/non-config types (e.g. int, list[str], a non-pydantic dataclass), the config parameter is valid.
If you need different config without editing the model, create a subclass with the desired model_config and adapt that.

Example fix

# before
TypeAdapter(MyModel, config=ConfigDict(strict=True))
# after
class StrictModel(MyModel):
    model_config = ConfigDict(strict=True)
TypeAdapter(StrictModel)

Defensive patterns

Strategy: validation

Validate before calling

from pydantic._internal import _model_construction
from pydantic import BaseModel, TypeAdapter

def adapter_for(tp, config=None):
    from pydantic.type_adapter import _type_has_config
    if config is not None and _type_has_config(tp):
        raise ValueError(f'{tp!r} carries its own config; set model_config on the type instead')
    return TypeAdapter(tp, config=config)

Type guard

from pydantic.type_adapter import _type_has_config
def type_accepts_adapter_config(tp) -> bool:
    return not _type_has_config(tp)

Prevention

Set config on the model/dataclass/TypedDict itself, not on the TypeAdapter wrapping it.
Only pass config to TypeAdapter for plain/non-config types.
Create a subclass with the desired model_config if you need a different config without editing the original.

When it happens

Trigger: TypeAdapter(MyModel, config=ConfigDict(strict=True)) where MyModel subclasses BaseModel; TypeAdapter over a dataclass or TypedDict with a config kwarg.

Common situations: Trying to override a model's strict/coerce behavior at the adapter level; wrapping an existing model with extra validation config.

Related errors

AI-assisted analysis of pydantic/pydantic@2e5f0e2b42 (2026-08-04). Data as JSON: /data/errors/16ef1f82121c87e1.json. Report an issue: GitHub.