diff --git a/metagpt/ext/android_assistant/actions/parse_record.py b/metagpt/ext/android_assistant/actions/parse_record.py
index 304daf6556..1eec7bdc1e 100644
--- a/metagpt/ext/android_assistant/actions/parse_record.py
+++ b/metagpt/ext/android_assistant/actions/parse_record.py
@@ -109,7 +109,7 @@ async def run(self, task_dir: Path, docs_dir: Path):
 
                 logger.info(f"Waiting for GPT-4V to generate documentation for the element {resource_id}")
                 node = await RECORD_PARSE_NODE.fill(
-                    context=context, llm=self.llm, images=[img_before_base64, img_after_base64]
+                    req=context, llm=self.llm, images=[img_before_base64, img_after_base64]
                 )
                 if "error" in node.content:
                     return AndroidActionOutput(action_state=RunState.FAIL)
diff --git a/metagpt/ext/android_assistant/actions/screenshot_parse.py b/metagpt/ext/android_assistant/actions/screenshot_parse.py
index 4d8bb0e1eb..fd3f6a5ab7 100644
--- a/metagpt/ext/android_assistant/actions/screenshot_parse.py
+++ b/metagpt/ext/android_assistant/actions/screenshot_parse.py
@@ -143,7 +143,7 @@ async def run(
 
         ui_doc = self._makeup_ui_document(elem_list, docs_dir)
         context = parse_template.format(ui_document=ui_doc, task_description=task_desc, last_act=last_act)
-        node = await SCREENSHOT_PARSE_NODE.fill(context=context, llm=self.llm, images=[img_base64])
+        node = await SCREENSHOT_PARSE_NODE.fill(req=context, llm=self.llm, images=[img_base64])
 
         if "error" in node.content:
             return AndroidActionOutput(action_state=RunState.FAIL)
diff --git a/metagpt/ext/android_assistant/actions/self_learn_and_reflect.py b/metagpt/ext/android_assistant/actions/self_learn_and_reflect.py
index 5e9cfbb454..7c5cf05c80 100644
--- a/metagpt/ext/android_assistant/actions/self_learn_and_reflect.py
+++ b/metagpt/ext/android_assistant/actions/self_learn_and_reflect.py
@@ -101,7 +101,7 @@ async def run_self_learn(
         self_explore_template = screenshot_parse_self_explore_template
         context = self_explore_template.format(task_description=task_desc, last_act=last_act)
 
-        node = await SCREENSHOT_PARSE_NODE.fill(context=context, llm=self.llm, images=[img_base64])
+        node = await SCREENSHOT_PARSE_NODE.fill(req=context, llm=self.llm, images=[img_base64])
         logger.debug(f"fill result:{node}")
         if "error" in node.content:
             return AndroidActionOutput(action_state=RunState.FAIL)
diff --git a/metagpt/ext/android_assistant/roles/android_assistant.py b/metagpt/ext/android_assistant/roles/android_assistant.py
index 97d66d30e4..9f94b6a00e 100644
--- a/metagpt/ext/android_assistant/roles/android_assistant.py
+++ b/metagpt/ext/android_assistant/roles/android_assistant.py
@@ -96,7 +96,7 @@ async def _observe(self, ignore_memory=True) -> int:
             ignore_memory = False
             state_val = newest_msg.content.split(".")[-1]  # RoundCount: 1, action_state: RunState.SUCCESS
             logger.warning(f"Latest action_state is {state_val}, will run in the remainder rounds without `react`")
-        return await super()._observe(ignore_memory)
+        return await super()._observe()
 
     async def _act(self) -> Message:
         logger.info(f"{self._setting}: to do {self.rc.todo}({self.rc.todo.name})")