Update app.py

app.py

@@ -91,14 +91,20 @@ def respond_with_langchain(
         
         response = llm.invoke(prompt_text)
         
-        if len(response) > 3000:
-            response = response[:3000] + "... [Response truncated for length]"
+        # Ensure response is clean and properly limited
+        if isinstance(response, str):
+            response = response.strip()
+            # Conservative limit to prevent content-length issues
+            if len(response) > 2500:
+                response = response[:2500] + "... [Response truncated - ask for continuation]"
         
         full_response = f"*{mode}*\n\n{response}"
-        yield full_response
+        
+        # Return directly instead of yielding (no streaming)
+        return full_response
         
     except Exception as e:
-        yield f"Sorry, I encountered an error: {str(e)[:200]}"
+        return f"Sorry, I encountered an error: {str(e)[:150]}"
 
 # Create the Gradio interface
 demo = gr.ChatInterface(