diff --git a/setup.cfg b/setup.cfg
index e05f599067..946c462fb0 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -22,7 +22,7 @@ install_requires =
   requests ~= 2.25.0 # TODO We should use httpx instead to avoid a new dependency
   # Client
   pandas >=1.0.0,<2.0.0 # For data loading
-
+  pydantic ~= 1.8.1
 
 [options.packages.find]
 where = src
@@ -36,7 +36,6 @@ where = src
 server =
     # Basic dependencies
     fastapi ~= 0.63.0
-    pydantic ~= 1.7.0
     uvicorn[standard] ~= 0.13.4
     elasticsearch >= 7.1.0,<8.0.0
     smart-open
diff --git a/src/rubrix/client/models.py b/src/rubrix/client/models.py
index c34150ece9..c804186ba3 100644
--- a/src/rubrix/client/models.py
+++ b/src/rubrix/client/models.py
@@ -52,7 +52,7 @@ class TextClassificationRecord(BaseModel):
 
     Attributes
     ----------
-    inputs : `Dict[str, Any]`
+    inputs : `Dict[str, Union[str, List[str]]]`
         The inputs of the record
     prediction : `List[Tuple[str, float]]`, optional
         A list of tuples containing the predictions for the record. The first entry of the tuple is the predicted label,
@@ -80,7 +80,7 @@ class TextClassificationRecord(BaseModel):
         The timestamp of the record. Default: None
     """
 
-    inputs: Dict[str, Any]
+    inputs: Dict[str, Union[str, List[str]]]
 
     prediction: Optional[List[Tuple[str, float]]] = None
     annotation: Optional[Union[str, List[str]]] = None
diff --git a/src/rubrix/server/tasks/text_classification/api/model.py b/src/rubrix/server/tasks/text_classification/api/model.py
index 9d89ef37cf..4287614cff 100644
--- a/src/rubrix/server/tasks/text_classification/api/model.py
+++ b/src/rubrix/server/tasks/text_classification/api/model.py
@@ -81,7 +81,7 @@ class CreationTextClassificationRecord(BaseRecord[TextClassificationAnnotation])
     Attributes:
     -----------
 
-    inputs: Dict[str, Any]
+    inputs: Dict[str, Union[str, List[str]]]
         The input data text
 
     multi_label: bool
@@ -93,7 +93,7 @@ class CreationTextClassificationRecord(BaseRecord[TextClassificationAnnotation])
         The dictionary key must be aligned with provided record text. Optional
     """
 
-    inputs: Dict[str, Any]
+    inputs: Dict[str, Union[str, List[str]]]
     multi_label: bool = False
     explanation: Dict[str, List[TokenAttributions]] = None
 
diff --git a/tests/client/test_models.py b/tests/client/test_models.py
index f6983118e9..404f517804 100644
--- a/tests/client/test_models.py
+++ b/tests/client/test_models.py
@@ -1,4 +1,5 @@
 import pytest
+from pydantic import ValidationError
 
 from rubrix.client.models import TextClassificationRecord
 from rubrix.client.models import TokenClassificationRecord
@@ -36,3 +37,9 @@ def test_token_classification_record(annotation, status, expected_status):
         text="test text", tokens=["test", "text"], annotation=annotation, status=status
     )
     assert record.status == expected_status
+
+
+def test_text_classification_record_none_inputs():
+    """Test validation error for None in inputs"""
+    with pytest.raises(ValidationError):
+        TextClassificationRecord(inputs={"text": None})
\ No newline at end of file
diff --git a/tests/server/text_classification/test_model.py b/tests/server/text_classification/test_model.py
index f280768412..3b7b90fb3e 100644
--- a/tests/server/text_classification/test_model.py
+++ b/tests/server/text_classification/test_model.py
@@ -9,16 +9,6 @@
 )
 
 
-def test_flatten_inputs():
-    data = {
-        "inputs": {
-            "mail": {"subject": "The mail subject", "body": "This is a large text body"}
-        }
-    }
-    record = TextClassificationRecord.parse_obj(data)
-    assert list(record.inputs.keys()) == ["mail.subject", "mail.body"]
-
-
 def test_flatten_metadata():
     data = {
         "inputs": {"text": "bogh"},