diff --git a/letta/agent.py b/letta/agent.py
index 50264bc739..07740c1e34 100644
--- a/letta/agent.py
+++ b/letta/agent.py
@@ -535,7 +535,7 @@ def _get_ai_reply(
                 # agent_state=self.agent_state,
                 llm_config=self.agent_state.llm_config,
                 messages=message_sequence,
-                user_id=self.agent_state.user_id,
+                user_id=self.agent_state.id,
                 functions=allowed_functions,
                 functions_python=self.functions_python,
                 function_call=function_call,
diff --git a/letta/llm_api/anthropic.py b/letta/llm_api/anthropic.py
index 9df4cec248..97df09375b 100644
--- a/letta/llm_api/anthropic.py
+++ b/letta/llm_api/anthropic.py
@@ -360,5 +360,11 @@ def anthropic_chat_completions_request(
     data.pop("user", None)
     data.pop("tool_choice", None)
 
+    # insert user_id for Anthropic
+    if "metadata" not in data:
+        data["metadata"] = {}
+    if user_id is not None:
+        data["metadata"]["user_id"] = user_id
+
     response_json = make_post_request(url, headers, data)
     return convert_anthropic_response_to_chatcompletion(response_json=response_json, inner_thoughts_xml_tag=inner_thoughts_xml_tag)
diff --git a/letta/llm_api/llm_api_tools.py b/letta/llm_api/llm_api_tools.py
index 9a6374b511..278a6dcc3b 100644
--- a/letta/llm_api/llm_api_tools.py
+++ b/letta/llm_api/llm_api_tools.py
@@ -150,7 +150,15 @@ def create(
             # only is a problem if we are *not* using an openai proxy
             raise ValueError(f"OpenAI key is missing from letta config file")
 
-        data = build_openai_chat_completions_request(llm_config, messages, user_id, functions, function_call, use_tool_naming, max_tokens)
+        data = build_openai_chat_completions_request(
+            llm_config=llm_config,
+            messages=messages,
+            user_id=user_id,
+            functions=functions,
+            function_call=function_call,
+            use_tool_naming=use_tool_naming,
+            max_tokens=max_tokens,
+        )
         if stream:  # Client requested token streaming
             data.stream = True
             assert isinstance(stream_interface, AgentChunkStreamingInterface) or isinstance(
@@ -253,7 +261,7 @@ def create(
                 messages=[cast_message_to_subtype(m.to_openai_dict()) for m in messages],
                 tools=[{"type": "function", "function": f} for f in functions] if functions else None,
                 # tool_choice=function_call,
-                # user=str(user_id),
+                user=user_id,
                 # NOTE: max_tokens is required for Anthropic API
                 max_tokens=1024,  # TODO make dynamic
             ),
diff --git a/letta/memory.py b/letta/memory.py
index a873226e5d..0acb1176d6 100644
--- a/letta/memory.py
+++ b/letta/memory.py
@@ -81,7 +81,7 @@ def summarize_messages(
     llm_config_no_inner_thoughts.put_inner_thoughts_in_kwargs = False
     response = create(
         llm_config=llm_config_no_inner_thoughts,
-        user_id=agent_state.user_id,
+        user_id=agent_state.id,
         messages=message_sequence,
         stream=False,
     )