diff --git a/lumen/ai/app.py b/lumen/ai/app.py
index 7a66f89d..8fe1b32e 100644
--- a/lumen/ai/app.py
+++ b/lumen/ai/app.py
@@ -3,6 +3,7 @@
 import param
 
 from panel.config import config, panel_extension
+from panel.io.resources import CSS_URLS
 from panel.io.state import state
 from panel.layout import Row
 from panel.viewable import Viewer
@@ -55,6 +56,9 @@ class LumenAI(Viewer):
     llm = param.ClassSelector(class_=Llm, default=OpenAI(), doc="""
         The LLM provider to be used by default""")
 
+    show_controls = param.Boolean(default=True, doc="""
+        Whether to show assistant controls in the sidebar.""")
+
     template = param.Selector(
         default=config.param.template.names['fast'],
         objects=config.param.template.names, doc="""
@@ -121,6 +125,7 @@ def show(self, **kwargs):
         return self._create_view(server=True).show(**kwargs)
 
     def _create_view(self, server: bool | None = None):
+        config.css_files.append(CSS_URLS['font-awesome'])
         if (state.curdoc and state.curdoc.session_context) or server is True:
             panel_extension(
                 *{ext for agent in self._assistant.agents for ext in agent._extensions}, template=self.template
@@ -129,7 +134,8 @@ def _create_view(self, server: bool | None = None):
             template = state.template
             template.title = self.title
             template.main.append(self._assistant)
-            template.sidebar.append(self._assistant.controls())
+            if self.show_controls:
+                template.sidebar.append(self._assistant.controls())
             return template
         return super()._create_view()
 
@@ -140,6 +146,8 @@ def servable(self, title: str | None = None, **kwargs):
         return self._create_view().servable(title, **kwargs)
 
     def __panel__(self):
+        if not self.show_controls:
+            return self._assistant
         return Row(
             Row(self._assistant.controls(), max_width=300),
             self._assistant
diff --git a/lumen/ai/llm.py b/lumen/ai/llm.py
index 8da21a3e..b3b48068 100644
--- a/lumen/ai/llm.py
+++ b/lumen/ai/llm.py
@@ -29,6 +29,10 @@ class Llm(param.Parameterized):
     # Allows defining a dictionary of default models.
     model_kwargs = param.Dict(default={})
 
+    # Whether the LLM supports streaming of any kind
+    _supports_stream = True
+
+    # Whether the LLM supports streaming of Pydantic model output
     _supports_model_stream = True
 
     __abstract = True
@@ -94,7 +98,8 @@ async def stream(
         model_key: str = "default",
         **kwargs,
     ):
-        if response_model and not self._supports_model_stream:
+        if ((response_model and not self._supports_model_stream) or
+            not self._supports_stream):
             yield await self.invoke(
                 messages,
                 system=system,
diff --git a/lumen/command/ai.py b/lumen/command/ai.py
index 662654d2..07ab09eb 100644
--- a/lumen/command/ai.py
+++ b/lumen/command/ai.py
@@ -13,7 +13,7 @@
 SOURCE_CODE = """
 import lumen.ai as lmai
 
-lmai.LumenAI({tables}).servable()"""
+lmai.LumenAI([{tables}]).servable()"""
 
 VALID_EXTENSIONS = ['.parq', '.parquet', '.csv', '.json']