diff --git a/Makefile b/Makefile
index 860d660..b1ba115 100644
--- a/Makefile
+++ b/Makefile
@@ -2,7 +2,7 @@
 
 setup:
 	docker compose down
-	docker compose up -d
+	docker compose up -d --build
 
 enter_container:
 	docker exec -it python_app bash
diff --git a/src/__init__.py b/src/__init__.py
index 59723e2..1aec526 100644
--- a/src/__init__.py
+++ b/src/__init__.py
@@ -1,9 +1,4 @@
-
-import logging
-
 # import sentry_sdk
-from common import settings
-from sentry_sdk.integrations.logging import LoggingIntegration
 
 
 # sentry_logging = LoggingIntegration(level=logging.INFO, event_level=logging.ERROR)
@@ -13,4 +8,4 @@
 #         integrations=[sentry_logging],
 #         environment=settings.ENV,
 #         traces_sample_rate=1.0,
-#     )
\ No newline at end of file
+#     )
diff --git a/src/common/__init__.py b/src/common/__init__.py
index 16ee946..33fdb66 100644
--- a/src/common/__init__.py
+++ b/src/common/__init__.py
@@ -9,7 +9,5 @@
 
 settings = read_yaml(os.path.join(base_dir, f"src/common/yaml_configs/{env}.yaml"))
 
-init_logger(
-    os.path.join(settings.BASE_DIR, "src/common/logger/logging_config.yaml")
-)
+init_logger(os.path.join(settings.BASE_DIR, "src/common/logger/logging_config.yaml"))
 app_logger = logging.getLogger(__name__)
diff --git a/src/common/config.py b/src/common/config.py
index b1c21b7..a318d01 100644
--- a/src/common/config.py
+++ b/src/common/config.py
@@ -1,5 +1,3 @@
-import os
-
 import yaml
 from pydantic import HttpUrl
 from pydantic_settings import BaseSettings
diff --git a/src/common/yaml_configs/local.yaml b/src/common/yaml_configs/local.yaml
index 756fbad..e1c1d1f 100644
--- a/src/common/yaml_configs/local.yaml
+++ b/src/common/yaml_configs/local.yaml
@@ -1,3 +1,3 @@
 ENV: "local"
 SENTRY_DSN:
-OPENAI_API_KEY: your-api-key
\ No newline at end of file
+OPENAI_API_KEY: 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
\ No newline at end of file
diff --git a/src/main.py b/src/main.py
index 5e4941b..072eb48 100644
--- a/src/main.py
+++ b/src/main.py
@@ -1,32 +1,36 @@
+import urllib.parse
+import uuid
+from typing import Any
+
 import streamlit as st
 import streamlit_mermaid as stmd
-import uuid
-from src.model.models import EnglishWord
-from src.common import app_logger
+from pydantic import ValidationError
+from service import OpenAISSEClient, WordQuestionBuilder
 
-from service import WordQuestionBuilder, OpenAISSEClient
+from src.common import app_logger
+from src.model.models import EnglishWord
 
 USER_NAME = "user"
 ASSISTANT_NAME = "assistant"
 
 
-def init_page():
+def init_page() -> None:
     st.set_page_config(
-        page_title="Utilizer: Boost your English active vocabulary 📚",
-        page_icon="📚"
+        page_title="UtiAIzer: Boost your English active vocabulary 📚", page_icon="📚"
     )
-    st.header("Utilizer: Boost your English active vocabulary")
+    st.header("UtiAIzer: Boost your English active vocabulary")
+    st.write(":orange[Generated by chatgpt. No responsibility for any content.]")
 
-
-def generate_session_id():
+def generate_session_id() -> str:
     return str(uuid.uuid4())
 
 
-def convart_to_url_part(sentence):
-    return sentence.replace(" ", "%20")
+def convart_to_url_part(sentence: str) -> str:
+    return urllib.parse.quote(sentence)
+
 
-def convart_to_under_score(sentence):
-    converted =  sentence.replace(" ", "_")
+def convart_to_under_score(sentence: str) -> str:
+    converted = sentence.replace(" ", "_")
     # del the fisrt, last char if it is "_"
     if converted[0] == "_":
         converted = converted[1:]
@@ -34,38 +38,51 @@ def convart_to_under_score(sentence):
         converted = converted[:-1]
     return converted
 
-def build_elsa_link(user_msg):
-    return f"\n - [Go To ELSA：{user_msg}](https://elsaspeak.com/en/learn-english/how-to-pronounce/{convart_to_url_part(user_msg)})"
 
+def build_elsa_link(user_msg: str) -> str:
+    return f"\n - [Go To ELSA:{user_msg}](https://elsaspeak.com/en/learn-english/how-to-pronounce/{convart_to_url_part(user_msg)})"
 
-def build_youglish_link(user_msg):
-    return f"\n\n - [Go To Youglish：{user_msg}](https://youglish.com/pronounce/{convart_to_url_part(user_msg)}/english?)"
 
+def build_youglish_link(user_msg: str) -> str:
+    return f"\n\n - [Go To Youglish:{user_msg}](https://youglish.com/pronounce/{convart_to_url_part(user_msg)}/english?)"
 
-def build_mermaid_graph_str(user_msg, result_str):
 
+def build_mermaid_graph_str(user_msg: str, result_str: str) -> str:
+    app_logger.info(f"result_str: {result_str}")
     li = map(convart_to_under_score, result_str.split("|"))
-    # skip the first element
+    # skip the first and second
+    next(li)
     next(li)
 
     code = f"""
     graph TD
-        {user_msg} --> {next(li)}
-        {user_msg} --> {next(li)}
-        {user_msg} --> {next(li)}
-        {user_msg} --> {next(li)}
+        {convart_to_under_score(user_msg)} --> {next(li)}
+        {convart_to_under_score(user_msg)} --> {next(li)}
+        {convart_to_under_score(user_msg)} --> {next(li)}
+        {convart_to_under_score(user_msg)} --> {next(li)}
     """
     return code
 
 
-def ask_llm_sse(key, user_msg, assistant_msg, message_placeholder, is_collocation=False):
+def ask_llm_sse(
+    key: str,
+    user_msg: str,
+    assistant_msg: str,
+    message_placeholder: Any,
+    is_collocation: bool = False,
+) -> tuple[str, str]:
     question = WordQuestionBuilder.call(key=key, word=EnglishWord(value=user_msg))
     answers = OpenAISSEClient.call(question)
 
     result_str = ""
+    count = 0
     for answer in answers:
+        if count == 60:
+            assistant_msg += "\n\n :red[Too many answers.] \n\n"
+            message_placeholder.write(assistant_msg)
+            break
 
-        if answer.value == '[END]':
+        if answer.value == "[END]":
             message_placeholder.write(assistant_msg)
             break
 
@@ -74,55 +91,107 @@ def ask_llm_sse(key, user_msg, assistant_msg, message_placeholder, is_collocatio
             message_placeholder.write(assistant_msg + "▌")
 
         result_str += answer.value
+        count += 1
 
     return assistant_msg, result_str
 
 
-def ask_several_questions_in_order(user_msg, assistant_msg, message_placeholder):
+def ask_several_questions_in_order(
+    user_msg: str, assistant_msg: str, message_placeholder: Any
+) -> str:
     # Meaning
-    assistant_msg += f"\n\n #### Meaning \n\n"
-    assistant_msg, _ = ask_llm_sse(key="meaning", user_msg=user_msg, assistant_msg=assistant_msg,
-                                   message_placeholder=message_placeholder)
+    assistant_msg += "\n\n #### Meaning \n\n"
+    assistant_msg, _ = ask_llm_sse(
+        key="meaning",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
     # Pronunciation
-    assistant_msg += f"\n\n #### Pronunciation \n\n"
+    assistant_msg += "\n\n #### Pronunciation \n\n"
     assistant_msg += "\n - IPA: "
-    assistant_msg, _ = ask_llm_sse(key="pronunciation", user_msg=user_msg, assistant_msg=assistant_msg,
-                                   message_placeholder=message_placeholder)
+    assistant_msg, _ = ask_llm_sse(
+        key="pronunciation",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
     # pronunciation_tip
     assistant_msg += "\n - Pronunciation Tip: "
-    assistant_msg, _ = ask_llm_sse(key="pronunciation_tip", user_msg=user_msg, assistant_msg=assistant_msg,
-                                   message_placeholder=message_placeholder)
+    assistant_msg, _ = ask_llm_sse(
+        key="pronunciation_tip",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
     # st.write(build_youglish_link(user_msg)
     assistant_msg += build_youglish_link(user_msg)
     assistant_msg += build_elsa_link(user_msg)
     message_placeholder.write(assistant_msg)
 
-    # Origin
-    assistant_msg += f"\n\n #### Origin \n\n"
-    assistant_msg, _ = ask_llm_sse(key="origin", user_msg=user_msg, assistant_msg=assistant_msg,
-                                   message_placeholder=message_placeholder)
+    # # Origin
+    # assistant_msg += f"\n\n #### Origin \n\n"
+    # assistant_msg, _ = ask_llm_sse(key="origin", user_msg=user_msg, assistant_msg=assistant_msg,
+    #                                message_placeholder=message_placeholder)
     # Synonym
-    assistant_msg += f"\n\n #### Synonym \n\n"
-    assistant_msg, _ = ask_llm_sse(key="synonym", user_msg=user_msg, assistant_msg=assistant_msg,
-                                   message_placeholder=message_placeholder)
+    assistant_msg += "\n\n #### Synonym \n\n"
+    assistant_msg, _ = ask_llm_sse(
+        key="synonym",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
+
+    # Synonym
+    assistant_msg += "\n\n #### Antonym \n\n"
+    assistant_msg, _ = ask_llm_sse(
+        key="antonym",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
+
     # Example Sentence
-    assistant_msg += f"\n\n #### Example Sentence \n\n"
-    assistant_msg, result_str = ask_llm_sse(key="example_sentence", user_msg=user_msg, assistant_msg=assistant_msg,
-                                            message_placeholder=message_placeholder)
-    assistant_msg += build_elsa_link(result_str)
+    assistant_msg += "\n\n #### Example Sentence \n\n"
+    assistant_msg, result_str = ask_llm_sse(
+        key="example_sentence",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+    )
+
+    # if AI find the example sentence, then build the url
+    if "I'm sorry" not in result_str and "Sorry" not in result_str:
+        assistant_msg += build_elsa_link(result_str)
+
+    # # making_sentence_tips
+    # assistant_msg += "\n\n #### Making Sentence Tips \n\n"
+    # assistant_msg, _ = ask_llm_sse(
+    #     key="making_sentence_tips",
+    #     user_msg=user_msg,
+    #     assistant_msg=assistant_msg,
+    #     message_placeholder=message_placeholder,
+    # )
 
     # Collocation
-    assistant_msg += f"\n\n #### Collocation MindMap \n\n"
-    assistant_msg, result_str = ask_llm_sse(key="collocation", user_msg=user_msg, assistant_msg=assistant_msg,
-                                             message_placeholder=message_placeholder, is_collocation=True)
+    assistant_msg += "\n\n #### Collocation MindMap \n\n"
+    assistant_msg, result_str = ask_llm_sse(
+        key="collocation",
+        user_msg=user_msg,
+        assistant_msg=assistant_msg,
+        message_placeholder=message_placeholder,
+        is_collocation=True,
+    )
     app_logger.info(f"result_str: {result_str}")
     code = build_mermaid_graph_str(user_msg, result_str)
     assistant_msg += code
     app_logger.info(f"code: {code}")
     stmd.st_mermaid(code)
+
     return assistant_msg
 
-def ask():
+
+def ask() -> None:
     current_session = st.session_state.current_session
     chat_log = st.session_state.chat_sessions[current_session]
     # 現在のセッションのチャット履歴を表示
@@ -139,13 +208,23 @@ def ask():
         with st.chat_message(ASSISTANT_NAME):
             message_placeholder = st.empty()
             assistant_msg = ""
-            assistant_msg = ask_several_questions_in_order(user_msg, assistant_msg, message_placeholder)
+            try:
+                assistant_msg = ask_several_questions_in_order(
+                    user_msg, assistant_msg, message_placeholder
+                )
+            except ValidationError:
+                st.error(
+                    "Please enter english word or phrase that is 30 characters or less including spaces."
+                )
+            except Exception as e:
+                app_logger.exception(e)
+                # st.error('Sorry, something went wrong.')
         # セッションにチャットログを追加
         chat_log.append({"name": USER_NAME, "msg": user_msg})
         chat_log.append({"name": ASSISTANT_NAME, "msg": assistant_msg})
 
 
-def main():
+def main() -> None:
     init_page()
 
     # セッション情報の初期化
@@ -158,7 +237,9 @@ def main():
 
     # Sidebarの実装
     session_list = list(st.session_state.chat_sessions.keys())
-    selected_session = st.sidebar.selectbox("Select Chat Session", session_list, index=len(session_list) - 1)
+    selected_session = st.sidebar.selectbox(
+        "Select Chat Session", session_list, index=len(session_list) - 1
+    )
     st.session_state.current_session = selected_session
 
     if st.sidebar.button("New Chat"):
@@ -169,5 +250,5 @@ def main():
     ask()
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     main()
diff --git a/src/model/models.py b/src/model/models.py
index f6c86a2..57c145e 100644
--- a/src/model/models.py
+++ b/src/model/models.py
@@ -1,15 +1,19 @@
 from pydantic import BaseModel, Field, validator
 
 
+class NotEnglishError(ValueError):
+    pass
+
+
 class EnglishWord(BaseModel):
-    value: str = Field(max_length=20)
+    value: str = Field(max_length=30)
 
     @validator("value")
-    def is_english(cls, v):
+    def is_english(cls, v: str) -> str:
         try:
-            v.encode('ascii')
+            v.encode("ascii")
         except UnicodeEncodeError:
-            raise ValueError("English only")
+            raise NotEnglishError("English only") from None
         return v
 
 
diff --git a/src/service/__init__.py b/src/service/__init__.py
index 8472eb1..35bddd4 100644
--- a/src/service/__init__.py
+++ b/src/service/__init__.py
@@ -1,7 +1,7 @@
-from .question_builder import WordQuestionBuilder
 from .openai_sse_client import OpenAISSEClient
+from .question_builder import WordQuestionBuilder
 
 __all__ = [
     "WordQuestionBuilder",
     "OpenAISSEClient",
-]
\ No newline at end of file
+]
diff --git a/src/service/openai_sse_client.py b/src/service/openai_sse_client.py
index 800db93..8cecfe8 100644
--- a/src/service/openai_sse_client.py
+++ b/src/service/openai_sse_client.py
@@ -1,17 +1,19 @@
+from collections.abc import Iterable
+
 import openai
-from src.model.models import Question, Answer
-from typing import Iterable
-from src.common import settings
-from src.common import app_logger
+
+from src.common import app_logger, settings
+from src.model.models import Answer, Question
 
 openai.api_key = settings.OPENAI_API_KEY
 
+
 class OpenAISSEClient:
     @classmethod
-    def call(cls, questions):
-        yield from cls(questions).__ask_llm_stream()
+    def call(cls, question: Question) -> Iterable[Answer]:
+        yield from cls(question).__ask_llm_stream()
 
-    def __init__(self, question) -> None:
+    def __init__(self, question: Question) -> None:
         self.__question = question
 
     def __ask_llm_stream(self) -> Iterable[Answer]:
@@ -20,22 +22,16 @@ def __ask_llm_stream(self) -> Iterable[Answer]:
             model="gpt-3.5-turbo",
             stream=True,  # SSEを使うための設定
             messages=[
-                {
-                    "role": "system",
-                    "content": "answer based on the following question"
-                },
-                {
-                    "role": "user",
-                    "content": f"{self.__question.value}"
-                }
+                {"role": "system", "content": "answer based on the following question"},
+                {"role": "user", "content": f"{self.__question.value}"},
             ],
         )
         app_logger.info(f"user_content: {self.__question.value}")
 
         for item in response:
             try:
-                content = item['choices'][0]['delta']['content']
-            except:
+                content = item["choices"][0]["delta"]["content"]
+            except Exception:
                 content = ""
             # dict型で返すことでよしなに変換してくれる
             yield Answer(value=content)
diff --git a/src/service/question_builder.py b/src/service/question_builder.py
index 1d657f8..035d530 100644
--- a/src/service/question_builder.py
+++ b/src/service/question_builder.py
@@ -1,31 +1,33 @@
-from src.model.models import Question, EnglishWord
+from src.model.models import EnglishWord, Question
 
 Key = str
 
 
 class QuestionBuilderBase:
-    templates = NotImplementedError
+    templates: dict[Key, str] = {}
 
     def __init__(self, key: str, word: EnglishWord) -> None:
         self._key = key
         self._word = word
 
     @classmethod
-    def call(cls, key: Key, word: EnglishWord):
+    def call(cls, key: Key, word: EnglishWord) -> Question:
         return cls(key=key, word=word).__make_question()
 
-    def __make_question(self):
+    def __make_question(self) -> Question:
         question_value = self.templates[self._key].format(word=self._word.value)
         return Question(value=question_value)
 
+
 class WordQuestionBuilder(QuestionBuilderBase):
     templates = {
-        "meaning": """output only the succinct meaning of "{word}" for kids""",
-        "origin": """output word roots of "{word}" for kids shortly""",
+        "meaning": """output the meaning of "{word}" shortly within 80 characters""",
+        "origin": """output word derivation of "{word}" within 100 characters""",
         "pronunciation": """output only the IPA for "{word}" within 20 characters""",
-        "pronunciation_tip": """Output the succinct pronunciation tip of "{word}" not based on IPA without preamble .Capitalize the characters should be emphasized.""",
-        "example_sentence": """Using "{word}, output one example sentence". Other words are easy for kids.""",
-        "making_sentence_tips": """output simple tips for making sentences with "{word}" correctly in terms on nuance and feeling for non-native speaker. Within 110 characters""",
-        "synonym": """output "{word}"s one synonym .Then tell the difference between "{word}" and the other for kids within 100 characters""",
-        "collocation" : """output "{word}"s collocations 6 times, separated with "|" . example : "apple pie|apple crisp" """,
+        "pronunciation_tip": """Output the pronunciation tip of "{word}" not based on IPA without preamble .Capitalize the characters should be emphasized. e.g: OHN-lee""",  # noqa
+        "example_sentence": """Using "{word}, output one example sentence within 70 characters".""",
+        "making_sentence_tips": """output tips for making sentences using "{word}" in terms on nuance and feeling for non-native speaker. Within 80 characters""",  # noqa
+        "synonym": """output "{word}"s one synonym .Then tell the difference between "{word}" and the other within 100 characters""",  # noqa
+        "antonym": """output "{word}"s one antonym .""",  # noqa
+        "collocation": """output "{word}"s collocations 7 times following grammer, separated with "|" . e.g: "{word} noun|{word} verb" """,  # noqa
     }
diff --git a/src/test_main.py b/src/test_main.py
deleted file mode 100644
index b5c9803..0000000
--- a/src/test_main.py
+++ /dev/null
@@ -1,10 +0,0 @@
-from service import WordQuestionBuilder, OpenAISSEClient
-from model.models import EnglishWord
-
-question = WordQuestionBuilder.call(key="collocation", word=EnglishWord(value="fart"))
-answers = OpenAISSEClient.call(question)
-result_str = ""
-for answer in answers:
-    result_str += answer.value
-
-breakpoint()
\ No newline at end of file
diff --git a/src/tmp/api.py b/src/tmp/api.py
deleted file mode 100644
index e49c7e1..0000000
--- a/src/tmp/api.py
+++ /dev/null
@@ -1,101 +0,0 @@
-import asyncio
-import json
-from functools import reduce
-
-import app.settings as settings
-import openai
-
-openai.api_key = settings.OPEN_AP_API_KEY
-
-
-class ChatGPTClient:
-    @classmethod
-    def async_call(cls, questions):
-        result_dict = asyncio.run(cls(questions).__async_call())
-        return result_dict
-
-    def __init__(self, questions) -> None:
-        self.__questions = questions
-
-    async def __async_call(self):
-        tasks = []
-        for question in self.__questions:
-            task = asyncio.create_task(self.__ask_chatgpt_async(question))
-            tasks.append(task)
-
-        results = await asyncio.gather(*tasks)
-        result_dict = reduce(lambda x, y: {**x, **y}, results)
-        return result_dict
-
-    @staticmethod
-    async def __ask_chatgpt_async(question):
-        response = openai.ChatCompletion.create(
-            model="gpt-3.5-turbo",
-            top_p=1,
-            temperature=0,
-            messages=[
-                {"role": "user", "content": question.text},
-            ],
-        )
-        return {question.key: response.choices[0]["message"]["content"].strip()}
-
-
-from dataclasses import dataclass
-
-
-@dataclass
-class Question:
-    key: str
-    text: str
-
-
-class QuestionBuilder:
-    templates = NotImplementedError
-
-    def __init__(self, **kwargs) -> None:
-        self.kwargs = kwargs
-
-    @classmethod
-    def call(cls, **kwargs):
-        return cls(**kwargs).__call()
-
-    def __call(self):
-        return [
-            Question(key, self.__make_question_async(key))
-            for key in self.templates.keys()
-        ]
-
-    def __make_question_async(self, key):
-        return self.templates[key].format(**self.kwargs)
-
-
-class WordQuestionBuilder(QuestionBuilder):
-    # WordQuestionは全てのこのワードが全部大事だしなー。
-
-    templates = {
-        "pronaunciation": """output only the IPA for "{word}" within 20 characters""",
-        "meaning_ja": """Output only the meaning for "{word}" within 20 characters.lang:ja""",
-        "meaning_en": """output only the succinct meaning of "{word}" for kids""",
-        "pronaunciation_tips": """Output the succinct pronunciation tip of "{word}" not based on IPA without preamble .Capitalize the characters should be emphasized.""",
-        "example_sentence": """Output one example sentence with "{word}". Other words are simple.""",
-        "making_sentence_tips": """output simple tips for making sentences with "{word}" correctly in terms on nuance and feeling for non-native speaker. Within 110 characters""",
-        "synonym": """output "{word}"s one synonym .Then tell the difference between "{word}" and the other for kids within 100 characters""",
-    }
-
-
-class SentenceFeedBackBuilder(QuestionBuilder):
-    templates = {
-        "grammar_correction": """correct grammar for "{sentence}" within 100 characters""",
-    }
-
-
-# Main program
-def main():
-    questions = WordQuestionBuilder.call(word="imprudent")
-    # questions = SentenceFeedBackBuilder.call(sentence="she no want me to buy the car.")
-    resp = ChatGPTCilent.async_call(questions)
-    print(resp)
-
-
-if __name__ == "__main__":
-    main()
\ No newline at end of file
diff --git a/src/tmp/parse.py b/src/tmp/parse.py
deleted file mode 100644
index a6c2540..0000000
--- a/src/tmp/parse.py
+++ /dev/null
@@ -1,14 +0,0 @@
-import re
-
-pronunciation = "uh-PRAY-zuhl"
-special_chars = "–,-, ,_"
-
-# Remove special characters except for space
-# 文字列内の特殊文字を', 'に置き換えるための変換テーブルを作成します
-trans = str.maketrans({char: "," for char in special_chars})
-
-# 文字列内の特殊文字を', 'に置き換えます
-output_str = pronunciation.translate(trans)
-
-# Join the pronunciation list with ", "
-print(output_str)
\ No newline at end of file
diff --git a/tests/module/test_temp_class.py b/tests/module/test_temp_class.py
deleted file mode 100644
index 4f23b22..0000000
--- a/tests/module/test_temp_class.py
+++ /dev/null
@@ -1,8 +0,0 @@
-from src.service.temp_class import TempClass
-
-
-class TestTempClass:
-    def test_add_temp(self) -> None:
-        got = TempClass().add_temp("test_input_str_")
-
-        assert "test_input_str_temp" == got