1 files changed, 9 insertions, 5 deletions

diff --git a/g4f/Provider/Vercel.py b/g4f/Provider/Vercel.py
index 2d20ca6a..2d856664 100644
--- a/g4f/Provider/Vercel.py
+++ b/g4f/Provider/Vercel.py
@@ -2,7 +2,7 @@ from __future__ import annotations
 
 import json, base64, requests, execjs, random, uuid
 
-from ..typing       import Any, TypedDict, CreateResult
+from ..typing       import Messages, TypedDict, CreateResult
 from .base_provider import BaseProvider
 from abc            import abstractmethod
 
@@ -17,8 +17,9 @@ class Vercel(BaseProvider):
     @abstractmethod
     def create_completion(
         model: str,
-        messages: list[dict[str, str]],
+        messages: Messages,
         stream: bool,
+        proxy: str = None,
         **kwargs
     ) -> CreateResult:
         if not model:
@@ -52,15 +53,18 @@ class Vercel(BaseProvider):
             'model'       : model_info[model]['id'],
             'messages'    : messages,
             'playgroundId': str(uuid.uuid4()),
-            'chatIndex'   : 0} | model_info[model]['default_params']
+            'chatIndex'   : 0,
+            **model_info[model]['default_params'],
+            **kwargs
+        }
 
         max_retries  = kwargs.get('max_retries', 20)
         for i in range(max_retries):
             response = requests.post('https://sdk.vercel.ai/api/generate', 
-                                    headers=headers, json=json_data, stream=True)
+                                    headers=headers, json=json_data, stream=True, proxies={"https": proxy})
             try:
                 response.raise_for_status()
-            except:
+            except Exception:
                 continue
             for token in response.iter_content(chunk_size=None):
                 yield token.decode()