diff --git a/Writer/Interface/Wrapper.py b/Writer/Interface/Wrapper.py
index 89b93035..14a19880 100644
--- a/Writer/Interface/Wrapper.py
+++ b/Writer/Interface/Wrapper.py
@@ -39,7 +39,9 @@ def LoadModels(self, Models: list):
                 Provider, ProviderModel, ModelHost, ModelOptions = (
                     self.GetModelAndProvider(Model)
                 )
-                print(f"DEBUG: Loading Model {ProviderModel} from {Provider}@{ModelHost}")
+                print(
+                    f"DEBUG: Loading Model {ProviderModel} from {Provider}@{ModelHost}"
+                )
 
                 if Provider == "ollama":
                     # Get ollama models (only once)
@@ -96,7 +98,30 @@ def LoadModels(self, Models: list):
                     )
 
                 elif Provider == "openai":
-                    raise NotImplementedError("OpenAI API not supported")
+                    # Validate OpenAI API Key
+                    if (
+                        not "OPENAI_API_KEY" in os.environ
+                        or os.environ["OPENAI_API_KEY"] == ""
+                    ):
+                        raise Exception(
+                            "OPENAI_API_KEY not found in environment variables. Add dummy if using local models"
+                        )
+                    self.ensure_package_is_installed("openai")
+                    import openai
+
+                    if ModelHost and not "://" in ModelHost:
+                        raise ValueError(
+                            f"Invalid Model Host URL {ModelHost}. Make sure to include the protocol (http/https)"
+                        )
+
+                    self.Clients[Model] = openai.OpenAI(
+                        api_key=os.environ["OPENAI_API_KEY"],
+                        base_url=(
+                            "https://api.openai.com/v1"
+                            if ModelHost is None
+                            else ModelHost
+                        ),
+                    )
 
                 elif Provider == "openrouter":
                     if (
@@ -126,23 +151,27 @@ def SafeGenerateText(
         _Model: str,
         _SeedOverride: int = -1,
         _Format: str = None,
-        _MinWordCount: int = 1
-        ):
+        _MinWordCount: int = 1,
+    ):
         """
         This function guarantees that the output will not be whitespace.
         """
 
-        NewMsg = self.ChatAndStreamResponse(_Logger, _Messages, _Model, _SeedOverride, _Format)
+        NewMsg = self.ChatAndStreamResponse(
+            _Logger, _Messages, _Model, _SeedOverride, _Format
+        )
 
-        while (self.GetLastMessageText(NewMsg).isspace()) or (len(self.GetLastMessageText(NewMsg).split(" ")) < _MinWordCount):
+        while (self.GetLastMessageText(NewMsg).isspace()) or (
+            len(self.GetLastMessageText(NewMsg).split(" ")) < _MinWordCount
+        ):
             _Logger.Log("Generation Failed, Reattempting Output", 7)
-            del _Messages[-1] # Remove failed attempt
-            NewMsg = self.ChatAndStreamResponse(_Logger, _Messages, _Model, random.randint(0, 99999), _Format)
+            del _Messages[-1]  # Remove failed attempt
+            NewMsg = self.ChatAndStreamResponse(
+                _Logger, _Messages, _Model, random.randint(0, 99999), _Format
+            )
 
         return NewMsg
 
-
-
     def ChatAndStreamResponse(
         self,
         _Logger,
@@ -310,7 +339,30 @@ def ChatAndStreamResponse(
                     m["role"] = "assistant"
 
         elif Provider == "openai":
-            raise NotImplementedError("OpenAI API not supported")
+
+            while True:
+                try:
+                    Stream = self.Clients[_Model].chat.completions.create(
+                        messages=_Messages,
+                        stream=True,
+                        model=ProviderModel,
+                    )
+                    _Messages.append(self.StreamResponse(Stream, Provider))
+                    break
+                except Exception as e:
+                    if MaxRetries > 0:
+                        _Logger.Log(
+                            f"Exception During Generation '{e}', {MaxRetries} Retries Remaining",
+                            7,
+                        )
+                        MaxRetries -= 1
+                    else:
+                        _Logger.Log(
+                            f"Max Retries Exceeded During Generation, Aborting!", 7
+                        )
+                        raise Exception(
+                            "Generation Failed, Max Retires Exceeded, Aborting"
+                        )
 
         elif Provider == "openrouter":
 
@@ -375,6 +427,8 @@ def StreamResponse(self, _Stream, _Provider: str):
                 ChunkText = chunk["message"]["content"]
             elif _Provider == "google":
                 ChunkText = chunk.text
+            elif _Provider == "openai":
+                ChunkText = chunk.choices[0].delta.content or ""
             else:
                 raise ValueError(f"Unsupported provider: {_Provider}")
 
@@ -405,23 +459,25 @@ def GetModelAndProvider(self, _Model: str):
             print(parsed)
             Provider = parsed.scheme
 
-            if "@" in parsed.netloc:
-                Model, Host = parsed.netloc.split("@")
+            FullPath = parsed.netloc + parsed.path
+
+            if "@" in FullPath:
+                Model, Host = FullPath.split("@")
 
             elif Provider == "openrouter":
-                Model = f"{parsed.netloc}{parsed.path}"
+                Model = FullPath
                 Host = None
 
             elif "ollama" in _Model:
                 if "@" in parsed.path:
-                   Model = parsed.netloc + parsed.path.split("@")[0]
-                   Host = parsed.path.split("@")[1]
+                    Model = FullPath.split("@")[0]
+                    Host = FullPath.split("@")[1]
                 else:
-                    Model = parsed.netloc
+                    Model = FullPath
                     Host = "localhost:11434"
 
             else:
-                Model = parsed.netloc
+                Model = FullPath
                 Host = None
             QueryParams = parse_qs(parsed.query)
 
diff --git a/requirements.txt b/requirements.txt
index 2b8ef125..c0965fc6 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,4 +1,4 @@
 ollama
 termcolor
 google.generativeai
-dotenv-python
\ No newline at end of file
+python-dotenv
\ No newline at end of file