Removed extra "Thinking..." and previous message getting repeated

app.py

@@ -29,8 +29,17 @@ st.sidebar.markdown(
 if "chat_history" not in st.session_state:
     st.session_state.chat_history = []
 
+# ⚠️ Avoid rendering the last assistant message prematurely
+history_to_render = st.session_state.chat_history.copy()
+if (
+    history_to_render
+    and history_to_render[-1]["role"] == "assistant"
+    and st.session_state.get("awaiting_response", False)
+):
+    history_to_render = history_to_render[:-1]
+
 # Display past chat messages
-for msg in st.session_state.chat_history:
+for msg in history_to_render:
     with st.chat_message(msg["role"]):
         st.markdown(msg["content"])
 
@@ -81,24 +90,29 @@ if user_input:
     if not asi_api_key:
         st.error("Please enter your API key in the sidebar.")
     else:
+        # Mark that we're about to stream a new response
+        st.session_state["awaiting_response"] = True
+
         # Add user message to chat history
         st.session_state.chat_history.append({"role": "user", "content": user_input})
         with st.chat_message("user"):
             st.markdown(user_input)
 
         with st.chat_message("assistant"):
-            with st.spinner("Thinking..."):
-                response_stream = stream_asi1_api(
-                    asi_api_key, st.session_state.chat_history
-                )
-                placeholder = st.empty()
-                partial_text = ""
-
-                for chunk in response_stream:
-                    partial_text += chunk
-                    placeholder.markdown(partial_text + "▌")
-
-                placeholder.markdown(partial_text)
-                st.session_state.chat_history.append(
-                    {"role": "assistant", "content": partial_text}
-                )
+            response_stream = stream_asi1_api(
+                asi_api_key, st.session_state.chat_history
+            )
+            placeholder = st.empty()
+            partial_text = ""
+
+            for chunk in response_stream:
+                partial_text += chunk
+                placeholder.markdown(partial_text + "▌")
+
+            placeholder.markdown(partial_text)
+
+        # Save final assistant message to history
+        st.session_state.chat_history.append(
+            {"role": "assistant", "content": partial_text}
+        )
+        st.session_state["awaiting_response"] = False