diff --git a/src/dolores.py b/src/dolores.py
index 1e4eacb..d8aa1fb 100644
--- a/src/dolores.py
+++ b/src/dolores.py
@@ -58,6 +58,7 @@ async def handle_mention(message):
         clean_message = message.clean_content.replace("@Dolores", "Dolores")
         clean_message = clean_message.replace("@everyone", "everyone")
         clean_message = clean_message.replace("@Testie", "Testie")
+        logger.info(f"Generating reply to following message: {clean_message}")
         reply = text_instance.generate_reply(clean_message)
     else:
         reply = "Hi"
diff --git a/src/modules/generation.py b/src/modules/generation.py
index 011ccb5..e933f89 100644
--- a/src/modules/generation.py
+++ b/src/modules/generation.py
@@ -48,19 +48,24 @@ def generate_reply(self, message):
             # Add the user's message to the message history
             message_history.append({"role": "user", "content": message})
 
-            # Generate a reply using the OpenAI API
-            response = openai.chat.completions.create(
-                model=os.environ["OPENAI_MODEL"],
-                messages=system_messages + list(message_history),
-                max_tokens=int(os.environ.get("MAX_TOKENS", 150)),
-                temperature=float(os.environ.get("TEMPERATURE", 0.9)),
-                top_p=float(os.environ.get("TOP_P", 1.0)),
-                frequency_penalty=float(os.environ.get("FREQUENCY_PENALTY", 0.0)),
-                presence_penalty=float(os.environ.get("PRESENCE_PENALTY", 0.6)),
-            )
-            reply = response.choices[0].message.content
-            # Add the reply to the message history
-            message_history.append({"role": "assistant", "content": reply})
+            try:
+                # Generate a reply using the OpenAI API
+                response = openai.chat.completions.create(
+                    model=os.environ["OPENAI_MODEL"],
+                    messages=system_messages + list(message_history),
+                    max_tokens=int(os.environ.get("MAX_TOKENS", 150)),
+                    temperature=float(os.environ.get("TEMPERATURE", 0.9)),
+                    top_p=float(os.environ.get("TOP_P", 1.0)),
+                    frequency_penalty=float(os.environ.get("FREQUENCY_PENALTY", 0.0)),
+                    presence_penalty=float(os.environ.get("PRESENCE_PENALTY", 0.6)),
+                )
+                reply = response.choices[0].message.content
+                logger.info(f"Reply generated: {reply}")
+                # Add the reply to the message history
+                message_history.append({"role": "assistant", "content": reply})
+            except Exception as e:
+                logger.error(f"Error generating reply: {e}")
+                logger.error(f"Messages: {system_messages + list(message_history)}")
 
         # Use a self-hosted LLM to generate a reply
         elif reply_method == "self":