diff --git a/clients/python/lorax/client.py b/clients/python/lorax/client.py
index fb88e0566..130b6c44e 100644
--- a/clients/python/lorax/client.py
+++ b/clients/python/lorax/client.py
@@ -3,7 +3,7 @@
 
 from aiohttp import ClientSession, ClientTimeout
 from pydantic import ValidationError
-from typing import Any, Dict, Optional, List, AsyncIterator, Iterator
+from typing import Any, Dict, Optional, List, AsyncIterator, Iterator, Union
 
 from lorax.types import (
     StreamResponse,
@@ -79,7 +79,7 @@ def generate(
         truncate: Optional[int] = None,
         typical_p: Optional[float] = None,
         watermark: bool = False,
-        response_format: Optional[ResponseFormat] = None,
+        response_format: Optional[Union[Dict[str, Any], ResponseFormat]] = None,
         decoder_input_details: bool = False,
     ) -> Response:
         """
@@ -125,7 +125,7 @@ def generate(
                 See [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666) for more information
             watermark (`bool`):
                 Watermarking with [A Watermark for Large Language Models](https://arxiv.org/abs/2301.10226)
-            response_format (`Optional[Dict[str, Any]]`):
+            response_format (`Optional[Union[Dict[str, Any], ResponseFormat]]`):
                 Optional specification of a format to impose upon the generated text, e.g.,:
                 ```
                 {
@@ -162,7 +162,7 @@ def generate(
             truncate=truncate,
             typical_p=typical_p,
             watermark=watermark,
-            response_format=response_format, # TODO: make object, ensure all variants of generate are migrated
+            response_format=response_format,
             decoder_input_details=decoder_input_details,
         )
         request = Request(inputs=prompt, stream=False, parameters=parameters)
@@ -175,9 +175,7 @@ def generate(
             timeout=self.timeout,
         )
 
-        # TODO: handle 422 errors
-        print(resp.text) # Failed to deserialize the JSON body into the target type: parameters.response_format: missing field `type` at line 1 column 854
-
+        # TODO: expose better error messages for 422 and similar errors
         payload = resp.json()
         if resp.status_code != 200:
             raise parse_error(resp.status_code, payload)
@@ -203,7 +201,7 @@ def generate_stream(
         truncate: Optional[int] = None,
         typical_p: Optional[float] = None,
         watermark: bool = False,
-        response_format: Optional[Dict[str, Any]] = None,
+        response_format: Optional[Union[Dict[str, Any], ResponseFormat]] = None,
     ) -> Iterator[StreamResponse]:
         """
         Given a prompt, generate the following stream of tokens
@@ -246,7 +244,7 @@ def generate_stream(
                 See [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666) for more information
             watermark (`bool`):
                 Watermarking with [A Watermark for Large Language Models](https://arxiv.org/abs/2301.10226)
-            response_format (`Optional[Dict[str, Any]]`):
+            response_format (`Optional[Union[Dict[str, Any], ResponseFormat]]`):
                 Optional specification of a format to impose upon the generated text, e.g.,:
                 ```
                 {
@@ -282,13 +280,13 @@ def generate_stream(
             truncate=truncate,
             typical_p=typical_p,
             watermark=watermark,
-            response_format=json.dumps(response_format) if response_format is not None else None,
+            response_format=response_format,
         )
         request = Request(inputs=prompt, stream=True, parameters=parameters)
 
         resp = requests.post(
             self.base_url,
-            json=request.dict(),
+            json=request.dict(by_alias=True),
             headers=self.headers,
             cookies=self.cookies,
             timeout=self.timeout,
@@ -384,7 +382,7 @@ async def generate(
         truncate: Optional[int] = None,
         typical_p: Optional[float] = None,
         watermark: bool = False,
-        response_format: Optional[Dict[str, Any]] = None,
+        response_format: Optional[Union[Dict[str, Any], ResponseFormat]] = None,
         decoder_input_details: bool = False,
     ) -> Response:
         """
@@ -430,7 +428,7 @@ async def generate(
                 See [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666) for more information
             watermark (`bool`):
                 Watermarking with [A Watermark for Large Language Models](https://arxiv.org/abs/2301.10226)
-            response_format (`Optional[Dict[str, Any]]`):
+            response_format (`Optional[Union[Dict[str, Any], ResponseFormat]]`):
                 Optional specification of a format to impose upon the generated text, e.g.,:
                 ```
                 {
@@ -468,14 +466,14 @@ async def generate(
             truncate=truncate,
             typical_p=typical_p,
             watermark=watermark,
-            response_format=json.dumps(response_format) if response_format is not None else None,
+            response_format=response_format,
         )
         request = Request(inputs=prompt, stream=False, parameters=parameters)
 
         async with ClientSession(
             headers=self.headers, cookies=self.cookies, timeout=self.timeout
         ) as session:
-            async with session.post(self.base_url, json=request.dict()) as resp:
+            async with session.post(self.base_url, json=request.dict(by_alias=True)) as resp:
                 payload = await resp.json()
 
                 if resp.status != 200:
@@ -501,7 +499,7 @@ async def generate_stream(
         truncate: Optional[int] = None,
         typical_p: Optional[float] = None,
         watermark: bool = False,
-        response_format: Optional[Dict[str, Any]] = None,
+        response_format: Optional[Union[Dict[str, Any], ResponseFormat]] = None,
     ) -> AsyncIterator[StreamResponse]:
         """
         Given a prompt, generate the following stream of tokens asynchronously
@@ -544,7 +542,7 @@ async def generate_stream(
                 See [Typical Decoding for Natural Language Generation](https://arxiv.org/abs/2202.00666) for more information
             watermark (`bool`):
                 Watermarking with [A Watermark for Large Language Models](https://arxiv.org/abs/2301.10226)
-            response_format (`Optional[Dict[str, Any]]`):
+            response_format (`Optional[Union[Dict[str, Any], ResponseFormat]]`):
                 Optional specification of a format to impose upon the generated text, e.g.,:
                 ```
                 {
@@ -580,14 +578,14 @@ async def generate_stream(
             truncate=truncate,
             typical_p=typical_p,
             watermark=watermark,
-            response_format=json.dumps(response_format) if response_format is not None else None,
+            response_format=response_format,
         )
         request = Request(inputs=prompt, stream=True, parameters=parameters)
 
         async with ClientSession(
             headers=self.headers, cookies=self.cookies, timeout=self.timeout
         ) as session:
-            async with session.post(self.base_url, json=request.dict()) as resp:
+            async with session.post(self.base_url, json=request.dict(by_alias=True)) as resp:
 
                 if resp.status != 200:
                     raise parse_error(resp.status, await resp.json())