diff --git a/spec/backend.spec b/spec/backend.spec
index 57d72f6f..786d7233 100644
--- a/spec/backend.spec
+++ b/spec/backend.spec
@@ -7,8 +7,8 @@ a = Analysis(
     binaries=[],
     datas=[
         ('./../src-python/models/overlay/fonts', 'fonts/'),
-        ('./../src-python/models/translation/prompt', 'prompt/'),
-        ('./../src-python/models/translation/languages', 'languages/'),
+        ('./../src-python/models/translation/translation_settings/prompt', 'translation_settings/prompt/'),
+        ('./../src-python/models/translation/translation_settings/languages', 'translation_settings/languages/'),
         ('./../.venv/Lib/site-packages/zeroconf', 'zeroconf/'),
         ('./../.venv/Lib/site-packages/openvr', 'openvr/'),
         ('./../.venv/Lib/site-packages/faster_whisper', 'faster_whisper/'),
diff --git a/spec/backend_cuda.spec b/spec/backend_cuda.spec
index 3d665926..0fc561f5 100644
--- a/spec/backend_cuda.spec
+++ b/spec/backend_cuda.spec
@@ -7,8 +7,8 @@ a = Analysis(
     binaries=[],
     datas=[
         ('./../src-python/models/overlay/fonts', 'fonts/'),
-        ('./../src-python/models/translation/prompt', 'prompt/'),
-        ('./../src-python/models/translation/languages', 'languages/'),
+        ('./../src-python/models/translation/translation_settings/prompt', 'prompt/'),
+        ('./../src-python/models/translation/translation_settings/languages', 'languages/'),
         ('./../.venv_cuda/Lib/site-packages/zeroconf', 'zeroconf/'),
         ('./../.venv_cuda/Lib/site-packages/openvr', 'openvr/'),
         ('./../.venv_cuda/Lib/site-packages/faster_whisper', 'faster_whisper/'),
diff --git a/src-python/controller.py b/src-python/controller.py
index ff91ea6e..67a5f1c1 100644
--- a/src-python/controller.py
+++ b/src-python/controller.py
@@ -292,6 +292,8 @@ class Controller:
                                 "data": None
                             },
                         )
+                    else:
+                        pass
                 except Exception as e:
                     # VRAM不足エラーの検出
                     is_vram_error, error_message = model.detectVRAMError(e)
@@ -412,6 +414,8 @@ class Controller:
                     translation_text = f" ({'/'.join(translation)})" if translation else ""
                     model.logger.info(f"[SENT] {message}{translation_text}")
 
+            model.addTranslationHistory("mic", message)
+
     def speakerMessage(self, result:dict) -> None:
         message = result["text"]
         language = result["language"]
@@ -452,6 +456,8 @@ class Controller:
                                 "data": None
                             },
                         )
+                    else:
+                        pass
                 except Exception as e:
                     # VRAM不足エラーの検出
                     is_vram_error, error_message = model.detectVRAMError(e)
@@ -594,6 +600,8 @@ class Controller:
                     translation_text = f" ({'/'.join(translation)})" if translation else ""
                     model.logger.info(f"[RECEIVED] {message}{translation_text}")
 
+            model.addTranslationHistory("speaker", message)
+
     def chatMessage(self, data) -> dict:
         id = data["id"]
         message = data["message"]
@@ -625,6 +633,8 @@ class Controller:
                                 "data": None
                             },
                         )
+                    else:
+                        pass
                 except Exception as e:
                     # VRAM不足エラーの検出
                     is_vram_error, error_message = model.detectVRAMError(e)
@@ -744,6 +754,8 @@ class Controller:
                 translation_text = f" ({'/'.join(translation)})" if translation else ""
                 model.logger.info(f"[CHAT] {message}{translation_text}")
 
+        model.addTranslationHistory("chat", message)
+
         return {
                 "status":200,
                 "result":{
diff --git a/src-python/docs/details/translation_prompt_history.md b/src-python/docs/details/translation_prompt_history.md
new file mode 100644
index 00000000..bf1ecd9b
--- /dev/null
+++ b/src-python/docs/details/translation_prompt_history.md
@@ -0,0 +1,186 @@
+# 翻訳プロンプトへの履歴注入（Chat/Mic/Speaker）
+
+LLM は直前までの会話文脈を理解して翻訳品質を向上させられます。そのため、システムプロンプトに最近の履歴（Chat/Mic/Speaker）を内包する機能を追加しました。大量のログでトークン消費が増えないよう、YAML 設定で取り込み範囲と上限を管理できます。
+
+## アーキテクチャ
+
+### 履歴管理（Model）
+
+**`model.py`** でChat/Mic/Speakerのメッセージ履歴を一元管理：
+
+```python
+# 履歴バッファ（最大50件）
+self.translation_history: list[dict] = []
+self.translation_history_max_items = 50
+
+# 履歴追加（オリジナルメッセージのみ、翻訳結果は保存しない）
+model.addTranslationHistory("chat", "こんにちは")
+model.addTranslationHistory("mic", "今日はいい天気")  
+model.addTranslationHistory("speaker", "Hello!")
+
+# 履歴取得
+history = model.getTranslationHistory(max_items=10)
+
+# 履歴クリア
+model.clearTranslationHistory()
+```
+
+### 自動注入（Model → Translator → 各LLMクライアント）
+
+- **`model.getTranslate()`** で履歴を取得し、`translator.translate(..., context_history=history)` に渡す。
+- **`Translator.translate()`** 側でエンジンごとの分岐直前に `setContextHistory()` を呼び、履歴をプロンプト組み立てに反映する。
+
+```python
+# model.getTranslate()
+history = self.getTranslationHistory()
+translation = self.translator.translate(
+  translator_name=translator_name,
+  weight_type=config.CTRANSLATE2_WEIGHT_TYPE,
+  source_language=source_language,
+  target_language=target_language,
+  target_country=target_country,
+  message=message,
+  context_history=history,
+)
+
+# Translator.translate() の一例（OpenAI）
+case "OpenAI_API":
+  if self.openai_client is None:
+    result = False
+  else:
+    if context_history:
+      self.openai_client.setContextHistory(context_history)
+    result = self.openai_client.translate(message, input_lang=source_language, output_lang=target_language)
+```
+
+### メッセージ処理（Controller）
+
+**`controller.py`** で各メッセージ処理完了後に履歴へ追加（翻訳の成否に関係なく、オリジナル文だけ保存）：
+
+- **Chat**: `chatMessage()` の末尾で `model.addTranslationHistory("chat", ...)`
+- **Mic**: `micMessage()` の末尾で `model.addTranslationHistory("mic", ...)`
+- **Speaker**: `speakerMessage()` の末尾で `model.addTranslationHistory("speaker", ...)`
+
+## 設定ファイル（例: OpenAI）
+
+`src-python/models/translation/translation_settings/prompt/translation_openai.yml`
+
+```yaml
+system_prompt: |
+  You are a helpful translation assistant.
+  Supported languages:
+  {supported_languages}
+
+  Translate the user provided text from {input_lang} to {output_lang}.
+  Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true               # 履歴をプロンプトへ注入するか
+  sources: [chat, mic, speaker]   # 取り込み対象の履歴種別
+  max_messages: 10                # 注入する履歴件数の上限（新しい順）
+  max_chars: 4000                 # 履歴整形後の最大文字数（超過時は先頭を切り捨て）
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
+```
+
+- `system_prompt`: 従来どおり、`{supported_languages}`/`{input_lang}`/`{output_lang}` が利用可能。
+- `history.use_history`: 履歴注入を有効化します。
+- `history.sources`: 取り込み対象ソース。`chat`/`mic`/`speaker` から選択。
+- `history.max_messages`: 新しい順に N 件を取り込みます。
+- `history.max_chars`: 整形後の履歴文字列の最大長。上限を超えた場合は先頭側を切り捨て（新しい文脈を優先）。
+- `history.header_template`: 履歴ヘッダの整形テンプレート。`{max_messages}`/`{history}` が利用可能。
+- `history.item_template`: 各履歴アイテムの整形テンプレート。`{timestamp}`（HH:MM形式）/`{source}`/`{text}` が利用可能。
+
+## 実装（OpenAI クライアント）
+
+`src-python/models/translation/translation_openai.py`
+
+- `OpenAIClient.setContextHistory(history_items: list[dict])` を追加。
+  - `history_items` は以下のキーを含む辞書の配列：
+    - `source`: `"chat" | "mic" | "speaker"`
+    - `text`:   文字列
+    - `timestamp`: ISO形式の日時文字列（HH:MM形式にフォーマットされてプロンプトに挿入）
+- `translate()` 呼び出し時、YAML の `history` 設定に基づき、指定履歴をシステムプロンプト末尾へ整形して注入します。
+- 文字数上限は簡易的に `max_chars` で制御（トークンカウントは行わず、過剰消費抑制用の安全策）。
+
+## 使い方
+
+### 基本的な流れ
+
+1. **メッセージ発生時に履歴追加**（controller.py で自動実行）
+```python
+# Chat送信時（オリジナルメッセージのみ保存）
+model.addTranslationHistory("chat", user_message)
+
+# Mic入力時（音声認識結果のみ保存）
+model.addTranslationHistory("mic", transcribed_text)
+
+# Speaker受信時（受信したオリジナルメッセージのみ保存）
+model.addTranslationHistory("speaker", received_text)
+```
+
+2. **翻訳時に自動注入**（model.py で自動実行）
+```python
+# getTranslate() 内で自動的に履歴が各LLMクライアントへ注入される
+translation = model.getTranslate(translator_name, ...)
+```
+
+3. **設定の調整**（YAML編集）
+```yaml
+history:
+  use_history: true          # 有効/無効
+  sources: [chat, mic]       # chatとmicのみ使う場合
+  max_messages: 5            # 最新5件のみ
+  max_chars: 2000            # 2000文字まで
+```
+
+### 手動で履歴を操作（必要な場合のみ）
+
+```python
+# 履歴をクリア
+model.clearTranslationHistory()
+
+# 履歴を取得
+recent_history = model.getTranslationHistory(max_items=10)
+
+# 手動で追加
+model.addTranslationHistory("chat", "カスタムメッセージ")
+```
+
+## 連携方法（開発者向け）
+
+既存のcontroller/model統合により、**自動で動作**します：
+
+1. ユーザーがChat入力 → `controller.chatMessage()` → メッセージ処理完了後に `model.addTranslationHistory()` 呼び出し（翻訳の成功/失敗に関係なく）
+2. マイク音声 → `controller.micMessage()` → メッセージ処理完了後に `model.addTranslationHistory()` 呼び出し（翻訳の成功/失敗に関係なく）
+3. スピーカー受信 → `controller.speakerMessage()` → メッセージ処理完了後に `model.addTranslationHistory()` 呼び出し（翻訳の成功/失敗に関係なく）
+4. 翻訳実行 → `model.getTranslate()` → LLMクライアントへ履歴を自動注入 → `client.translate()` で履歴付きプロンプト生成
+
+**重要**: 履歴にはオリジナルメッセージのみが保存されます。翻訳結果は履歴に含まれません。これによりトークン消費を抑え、文脈として必要な情報のみを提供します。
+
+**追加実装は不要です。** YAML設定を変更するだけで履歴注入の有効/無効や範囲を制御できます。
+
+## 連携方法
+
+## 対応状況
+
+✅ **全LLMクライアントに展開済み**
+
+以下のすべてのクライアントで履歴注入機能が利用可能です：
+- OpenAI (`translation_openai.py` / `translation_openai.yml`)
+- Gemini (`translation_gemini.py` / `translation_gemini.yml`)
+- Groq (`translation_groq.py` / `translation_groq.yml`)
+- OpenRouter (`translation_openrouter.py` / `translation_openrouter.yml`)
+- LMStudio (`translation_lmstudio.py` / `translation_lmstudio.yml`)
+- Ollama (`translation_ollama.py` / `translation_ollama.yml`)
+- Plamo (`translation_plamo.py` / `translation_plamo.yml`)
+
+各クライアントで同一の設定形式とAPIインターフェースを使用します：
+- `setContextHistory(history_items: list[dict])` メソッド
+- YAML の `history` セクション
+
+## 今後の拡張案
+
+- 実トークン見積りに基づく切り詰め（tiktoken 等）
+- 要約モデルを使った古い履歴の縮約
diff --git a/src-python/model.py b/src-python/model.py
index 81733b4e..335991ec 100644
--- a/src-python/model.py
+++ b/src-python/model.py
@@ -117,6 +117,8 @@ class Model:
         self.previous_receive_message = ""
         self.translator = Translator()
         self.keyword_processor = KeywordProcessor()
+        self.translation_history: list[dict] = []
+        self.translation_history_max_items = 20
         overlay_small_log_settings = copy.deepcopy(config.OVERLAY_SMALL_LOG_SETTINGS)
         overlay_large_log_settings = copy.deepcopy(config.OVERLAY_LARGE_LOG_SETTINGS)
         overlay_large_log_settings["ui_scaling"] = overlay_large_log_settings["ui_scaling"] * 0.25
@@ -378,16 +380,62 @@ class Model:
 
         return compatible_engines
 
+    def addTranslationHistory(self, source: str, text: str) -> None:
+        """Add a message to translation context history.
+        
+        Args:
+            source: "chat" | "mic" | "speaker"
+            text: message content
+        """
+        self.ensure_initialized()
+        if not text or not text.strip():
+            return
+        
+        history_item = {
+            "source": source,
+            "text": text.strip(),
+            "timestamp": datetime.now().isoformat(),
+        }
+        self.translation_history.append(history_item)
+        
+        # 最大件数を超えた場合は古いものを削除
+        if len(self.translation_history) > self.translation_history_max_items:
+            self.translation_history = self.translation_history[-self.translation_history_max_items:]
+    
+    def getTranslationHistory(self, max_items: int = None) -> list[dict]:
+        """Get recent translation context history.
+        
+        Args:
+            max_items: Maximum number of items to return (newest first)
+        
+        Returns:
+            List of history items
+        """
+        self.ensure_initialized()
+        if max_items is None or max_items <= 0:
+            return self.translation_history
+        return self.translation_history[-max_items:]
+    
+    def clearTranslationHistory(self) -> None:
+        """Clear all translation context history."""
+        self.ensure_initialized()
+        self.translation_history = []
+
     def getTranslate(self, translator_name, source_language, target_language, target_country, message):
         self.ensure_initialized()
         success_flag = False
+        
+        # Get context history for LLM-based translators
+        history = self.getTranslationHistory()
+        
         translation = self.translator.translate(
                         translator_name=translator_name,
                         weight_type=config.CTRANSLATE2_WEIGHT_TYPE,
                         source_language=source_language,
                         target_language=target_language,
                         target_country=target_country,
-                        message=message
+                        message=message,
+                        context_history=history
                 )
 
         # 翻訳失敗時のフェールセーフ処理
diff --git a/src-python/models/translation/prompt/translation_gemini.yml b/src-python/models/translation/prompt/translation_gemini.yml
deleted file mode 100644
index 8dbe4927..00000000
--- a/src-python/models/translation/prompt/translation_gemini.yml
+++ /dev/null
@@ -1,7 +0,0 @@
-system_prompt: |
-  You are a helpful translation assistant.
-  Supported languages:
-  {supported_languages}
-
-  Translate the user provided text from {input_lang} to {output_lang}.
-  Return ONLY the translated text. Do not add quotes or extra commentary.
\ No newline at end of file
diff --git a/src-python/models/translation/prompt/translation_lmstudio.yml b/src-python/models/translation/prompt/translation_lmstudio.yml
deleted file mode 100644
index 8dbe4927..00000000
--- a/src-python/models/translation/prompt/translation_lmstudio.yml
+++ /dev/null
@@ -1,7 +0,0 @@
-system_prompt: |
-  You are a helpful translation assistant.
-  Supported languages:
-  {supported_languages}
-
-  Translate the user provided text from {input_lang} to {output_lang}.
-  Return ONLY the translated text. Do not add quotes or extra commentary.
\ No newline at end of file
diff --git a/src-python/models/translation/prompt/translation_ollama.yml b/src-python/models/translation/prompt/translation_ollama.yml
deleted file mode 100644
index 8dbe4927..00000000
--- a/src-python/models/translation/prompt/translation_ollama.yml
+++ /dev/null
@@ -1,7 +0,0 @@
-system_prompt: |
-  You are a helpful translation assistant.
-  Supported languages:
-  {supported_languages}
-
-  Translate the user provided text from {input_lang} to {output_lang}.
-  Return ONLY the translated text. Do not add quotes or extra commentary.
\ No newline at end of file
diff --git a/src-python/models/translation/prompt/translation_openai.yml b/src-python/models/translation/prompt/translation_openai.yml
deleted file mode 100644
index 8dbe4927..00000000
--- a/src-python/models/translation/prompt/translation_openai.yml
+++ /dev/null
@@ -1,7 +0,0 @@
-system_prompt: |
-  You are a helpful translation assistant.
-  Supported languages:
-  {supported_languages}
-
-  Translate the user provided text from {input_lang} to {output_lang}.
-  Return ONLY the translated text. Do not add quotes or extra commentary.
\ No newline at end of file
diff --git a/src-python/models/translation/translation_gemini.py b/src-python/models/translation/translation_gemini.py
index 1c8a0161..8b68dc33 100644
--- a/src-python/models/translation/translation_gemini.py
+++ b/src-python/models/translation/translation_gemini.py
@@ -4,14 +4,14 @@ from langchain_google_genai import ChatGoogleGenerativeAI
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     print(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     sys.path.append(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     from translation_languages import translation_lang
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
 
 logger = logging.getLogger("langchain_google_genai")
 logger.setLevel(logging.ERROR)
@@ -57,9 +57,19 @@ class GeminiClient:
         self.model = None
 
         # プロンプト設定をYAMLファイルから読み込む
-        prompt_config = loadPromptConfig(root_path, "translation_gemini.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_gemini.yml")
         self.supported_languages = list(translation_lang["Gemini_API"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.gemini_llm = None
 
@@ -91,6 +101,16 @@ class GeminiClient:
             api_key=self.api_key,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
@@ -98,6 +118,41 @@ class GeminiClient:
             output_lang=output_lang
         )
 
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text}
diff --git a/src-python/models/translation/translation_groq.py b/src-python/models/translation/translation_groq.py
index 04c57dbe..3952f06f 100644
--- a/src-python/models/translation/translation_groq.py
+++ b/src-python/models/translation/translation_groq.py
@@ -4,13 +4,13 @@ from pydantic import SecretStr
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 def _authentication_check(api_key: str) -> bool:
@@ -73,9 +73,19 @@ class GroqClient:
         self.model = None
         self.base_url = "https://api.groq.com/openai/v1"
 
-        prompt_config = loadPromptConfig(root_path, "translation_groq.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_groq.yml")
         self.supported_languages = list(translation_lang["Groq_API"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.groq_llm = None
 
@@ -109,12 +119,58 @@ class GroqClient:
             streaming=False,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang,
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/translation_languages.py b/src-python/models/translation/translation_languages.py
index eccd5665..311bbd95 100644
--- a/src-python/models/translation/translation_languages.py
+++ b/src-python/models/translation/translation_languages.py
@@ -41,14 +41,14 @@ def _load_languages(path: str, filename: str) -> str:
     Returns:
         Absolute path to the resource file
     """
-    if os.path.exists(os.path.join(path, "_internal", "languages", "languages.yml")):
-        languages_path =  os.path.join(path, "_internal", "languages", "languages.yml")
-    elif os.path.exists(os.path.join(os.path.dirname(os.path.abspath(__file__)), "models", "translation", "languages", "languages.yml")):
-        languages_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "models", "translation", "languages", "languages.yml")
-    elif os.path.exists(os.path.join(os.path.dirname(os.path.abspath(__file__)), "languages", "languages.yml")):
-        languages_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "languages", "languages.yml")
+    if os.path.exists(os.path.join(path, "_internal", "translation_settings", "languages", filename)):
+        languages_path =  os.path.join(path, "_internal", "translation_settings", "languages", filename)
+    elif os.path.exists(os.path.join(os.path.dirname(os.path.abspath(__file__)), "models", "translation", "translation_settings", "languages", filename)):
+        languages_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "models", "translation", "translation_settings", "languages", filename)
+    elif os.path.exists(os.path.join(os.path.dirname(os.path.abspath(__file__)), "translation_settings", "languages", filename)):
+        languages_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), "translation_settings", "languages", filename)
     else:
-        raise FileNotFoundError(f"Prompt file not found: {filename}")
+        raise FileNotFoundError(f"Languages file not found: {filename}")
     with open(languages_path, "r", encoding="utf-8") as f:
         return yaml.safe_load(f)
 
@@ -99,7 +99,7 @@ def loadTranslationLanguages(path: str, force: bool = False) -> Dict[str, Any]:
         if _loaded and not force:
             return translation_lang
 
-        data = _load_languages(path, "languages/languages.yml")
+        data = _load_languages(path, "languages.yml")
 
         if not isinstance(data, dict):
             raise ValueError(
diff --git a/src-python/models/translation/translation_lmstudio.py b/src-python/models/translation/translation_lmstudio.py
index 16ff9545..0567fcad 100644
--- a/src-python/models/translation/translation_lmstudio.py
+++ b/src-python/models/translation/translation_lmstudio.py
@@ -4,13 +4,13 @@ import requests
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.abspath(__file__)))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 def _authentication_check(base_url: str | None = None) -> bool:
@@ -50,9 +50,19 @@ class LMStudioClient:
         self.model = None
         self.base_url = base_url  # None の場合は公式エンドポイント
 
-        prompt_config = loadPromptConfig(root_path, "translation_lmstudio.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_lmstudio.yml")
         self.supported_languages = list(translation_lang["LMStudio"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.openai_llm = None
 
@@ -86,12 +96,58 @@ class LMStudioClient:
             streaming=False,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang,
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/translation_ollama.py b/src-python/models/translation/translation_ollama.py
index 543e603a..8d1cae94 100644
--- a/src-python/models/translation/translation_ollama.py
+++ b/src-python/models/translation/translation_ollama.py
@@ -3,13 +3,13 @@ from langchain_ollama import ChatOllama
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.abspath(__file__)))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 def _authentication_check(base_url: str | None = None) -> bool:
@@ -48,9 +48,19 @@ class OllamaClient:
         self.model = None
         self.base_url = "http://localhost:11434"
 
-        prompt_config = loadPromptConfig(root_path, "translation_ollama.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_ollama.yml")
         self.supported_languages = list(translation_lang["Ollama"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.openai_llm = None
 
@@ -79,12 +89,58 @@ class OllamaClient:
             streaming=False,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang,
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/translation_openai.py b/src-python/models/translation/translation_openai.py
index cc09a90e..796de430 100644
--- a/src-python/models/translation/translation_openai.py
+++ b/src-python/models/translation/translation_openai.py
@@ -4,13 +4,13 @@ from pydantic import SecretStr
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 def _authentication_check(api_key: str, base_url: str | None = None) -> bool:
@@ -69,9 +69,19 @@ class OpenAIClient:
         self.model = None
         self.base_url = base_url  # None の場合は公式エンドポイント
 
-        prompt_config = loadPromptConfig(root_path, "translation_openai.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_openai.yml")
         self.supported_languages = list(translation_lang["OpenAI_API"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.openai_llm = None
 
@@ -105,12 +115,62 @@ class OpenAIClient:
             streaming=False,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang,
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            # filter by source and take newest N
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            # format items
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            # truncate by char limit to mitigate token use
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            # assemble header and append to system prompt
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/translation_openrouter.py b/src-python/models/translation/translation_openrouter.py
index cec7aee3..20746bbf 100644
--- a/src-python/models/translation/translation_openrouter.py
+++ b/src-python/models/translation/translation_openrouter.py
@@ -4,13 +4,13 @@ from pydantic import SecretStr
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 def _authentication_check(api_key: str) -> bool:
@@ -73,9 +73,19 @@ class OpenRouterClient:
         self.model = None
         self.base_url = "https://openrouter.ai/api/v1"
 
-        prompt_config = loadPromptConfig(root_path, "translation_openrouter.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_openrouter.yml")
         self.supported_languages = list(translation_lang["OpenRouter_API"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.openrouter_llm = None
 
@@ -109,12 +119,58 @@ class OpenRouterClient:
             streaming=False,
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang,
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/translation_plamo.py b/src-python/models/translation/translation_plamo.py
index 68d68754..1752f7fa 100644
--- a/src-python/models/translation/translation_plamo.py
+++ b/src-python/models/translation/translation_plamo.py
@@ -4,13 +4,13 @@ from pydantic import SecretStr
 
 try:
     from .translation_languages import translation_lang
-    from .translation_utils import loadPromptConfig
+    from .translation_utils import loadTranslatePromptConfig
 except Exception:
     import sys
     from os import path as os_path
     sys.path.append(os_path.dirname(os_path.dirname(os_path.dirname(os_path.abspath(__file__)))))
     from translation_languages import translation_lang, loadTranslationLanguages
-    from translation_utils import loadPromptConfig
+    from translation_utils import loadTranslatePromptConfig
     translation_lang = loadTranslationLanguages(path=".", force=True)
 
 BASE_URL = "https://api.platform.preferredai.jp/v1"
@@ -44,9 +44,19 @@ class PlamoClient:
         self.base_url = BASE_URL
         self.model = None
 
-        prompt_config = loadPromptConfig(root_path, "translation_plamo.yml")
+        prompt_config = loadTranslatePromptConfig(root_path, "translation_plamo.yml")
         self.supported_languages = list(translation_lang["Plamo_API"]["source"].keys())
         self.prompt_template = prompt_config["system_prompt"]
+        # history config (optional)
+        self.history_cfg = prompt_config.get("history", {
+            "use_history": False,
+            "sources": [],
+            "max_messages": 0,
+            "max_chars": 0,
+            "header_template": "",
+            "item_template": "[{source}] {role}: {text}",
+        })
+        self._context_history: list[dict] = []
 
         self.plamo_llm = None
 
@@ -80,12 +90,58 @@ class PlamoClient:
             api_key=SecretStr(self.api_key),
         )
 
+    def setContextHistory(self, history_items: list[dict]) -> None:
+        """Set recent conversation history for prompt injection.
+
+        Each item should be a dict containing:
+        - source: "chat" | "mic" | "speaker"
+        - text: message string
+        - timestamp: ISO format datetime string
+        """
+        self._context_history = history_items or []
+
     def translate(self, text: str, input_lang: str, output_lang: str) -> str:
         system_prompt = self.prompt_template.format(
             supported_languages=self.supported_languages,
             input_lang=input_lang,
             output_lang=output_lang
         )
+
+        # Inject recent conversation history if enabled by YAML config
+        if self.history_cfg.get("use_history"):
+            allowed_sources = set(self.history_cfg.get("sources", []))
+            max_messages = int(self.history_cfg.get("max_messages", 0))
+            max_chars = int(self.history_cfg.get("max_chars", 0))
+            item_tmpl = self.history_cfg.get("item_template", "[{source}] {role}: {text}")
+            header_tmpl = self.history_cfg.get("header_template", "{history}")
+
+            filtered = [h for h in self._context_history if h.get("source") in allowed_sources]
+            recent = filtered[-max_messages:] if max_messages > 0 else filtered
+            formatted_items = []
+            for h in recent:
+                # Format timestamp as HH:MM to save tokens
+                timestamp_str = ''
+                if 'timestamp' in h:
+                    from datetime import datetime
+                    try:
+                        ts = datetime.fromisoformat(h['timestamp'])
+                        timestamp_str = ts.strftime('%H:%M')
+                    except:
+                        timestamp_str = ''
+                formatted_items.append(
+                    item_tmpl.format(
+                        timestamp=timestamp_str,
+                        source=h.get("source", ""),
+                        text=h.get("text", ""),
+                    )
+                )
+            history_blob = "\n".join(formatted_items).strip()
+            if max_chars and len(history_blob) > max_chars:
+                history_blob = history_blob[-max_chars:]
+            history_header = header_tmpl.format(max_messages=max_messages, history=history_blob)
+            if history_header:
+                system_prompt = f"{system_prompt}\n\n{history_header}"
+
         messages = [
             {"role": "system", "content": system_prompt},
             {"role": "user", "content": text},
diff --git a/src-python/models/translation/languages/languages.yml b/src-python/models/translation/translation_settings/languages/languages.yml
similarity index 100%
rename from src-python/models/translation/languages/languages.yml
rename to src-python/models/translation/translation_settings/languages/languages.yml
diff --git a/src-python/models/translation/prompt/translation_groq.yml b/src-python/models/translation/translation_settings/prompt/translation_gemini.yml
similarity index 51%
rename from src-python/models/translation/prompt/translation_groq.yml
rename to src-python/models/translation/translation_settings/prompt/translation_gemini.yml
index bc256c3d..96e67a7e 100644
--- a/src-python/models/translation/prompt/translation_groq.yml
+++ b/src-python/models/translation/translation_settings/prompt/translation_gemini.yml
@@ -5,3 +5,12 @@ system_prompt: |
 
   Translate the user provided text from {input_lang} to {output_lang}.
   Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
\ No newline at end of file
diff --git a/src-python/models/translation/translation_settings/prompt/translation_groq.yml b/src-python/models/translation/translation_settings/prompt/translation_groq.yml
new file mode 100644
index 00000000..03d38844
--- /dev/null
+++ b/src-python/models/translation/translation_settings/prompt/translation_groq.yml
@@ -0,0 +1,16 @@
+system_prompt: |
+  You are a helpful translation assistant.
+  Supported languages:
+  {supported_languages}
+
+  Translate the user provided text from {input_lang} to {output_lang}.
+  Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
diff --git a/src-python/models/translation/prompt/translation_openrouter.yml b/src-python/models/translation/translation_settings/prompt/translation_lmstudio.yml
similarity index 51%
rename from src-python/models/translation/prompt/translation_openrouter.yml
rename to src-python/models/translation/translation_settings/prompt/translation_lmstudio.yml
index bc256c3d..96e67a7e 100644
--- a/src-python/models/translation/prompt/translation_openrouter.yml
+++ b/src-python/models/translation/translation_settings/prompt/translation_lmstudio.yml
@@ -5,3 +5,12 @@ system_prompt: |
 
   Translate the user provided text from {input_lang} to {output_lang}.
   Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
\ No newline at end of file
diff --git a/src-python/models/translation/translation_settings/prompt/translation_ollama.yml b/src-python/models/translation/translation_settings/prompt/translation_ollama.yml
new file mode 100644
index 00000000..96e67a7e
--- /dev/null
+++ b/src-python/models/translation/translation_settings/prompt/translation_ollama.yml
@@ -0,0 +1,16 @@
+system_prompt: |
+  You are a helpful translation assistant.
+  Supported languages:
+  {supported_languages}
+
+  Translate the user provided text from {input_lang} to {output_lang}.
+  Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
\ No newline at end of file
diff --git a/src-python/models/translation/translation_settings/prompt/translation_openai.yml b/src-python/models/translation/translation_settings/prompt/translation_openai.yml
new file mode 100644
index 00000000..9b71cfbb
--- /dev/null
+++ b/src-python/models/translation/translation_settings/prompt/translation_openai.yml
@@ -0,0 +1,16 @@
+system_prompt: |
+  You are a helpful translation assistant.
+  Supported languages:
+  {supported_languages}
+
+  Translate the user provided text from {input_lang} to {output_lang}.
+  Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker] # 取り込み対象の履歴種別
+  max_messages: 5               # 注入する履歴件数の上限（新しい順）
+  max_chars: 4000               # 履歴整形後の最大文字数（超過時は先頭を切り捨て）
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
\ No newline at end of file
diff --git a/src-python/models/translation/translation_settings/prompt/translation_openrouter.yml b/src-python/models/translation/translation_settings/prompt/translation_openrouter.yml
new file mode 100644
index 00000000..03d38844
--- /dev/null
+++ b/src-python/models/translation/translation_settings/prompt/translation_openrouter.yml
@@ -0,0 +1,16 @@
+system_prompt: |
+  You are a helpful translation assistant.
+  Supported languages:
+  {supported_languages}
+
+  Translate the user provided text from {input_lang} to {output_lang}.
+  Return ONLY the translated text. Do not add quotes or extra commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
diff --git a/src-python/models/translation/prompt/translation_plamo.yml b/src-python/models/translation/translation_settings/prompt/translation_plamo.yml
similarity index 55%
rename from src-python/models/translation/prompt/translation_plamo.yml
rename to src-python/models/translation/translation_settings/prompt/translation_plamo.yml
index c0afe533..a97a57b9 100644
--- a/src-python/models/translation/prompt/translation_plamo.yml
+++ b/src-python/models/translation/translation_settings/prompt/translation_plamo.yml
@@ -4,4 +4,13 @@ system_prompt: |
   {supported_languages}
 
   Translate the following text from {input_lang} to {output_lang}.
-  output only the translated text without any additional commentary.
\ No newline at end of file
+  output only the translated text without any additional commentary.
+history:
+  use_history: true
+  sources: [chat, mic, speaker]
+  max_messages: 5
+  max_chars: 4000
+  header_template: |
+    Conversation context (recent {max_messages} messages):
+    {history}
+  item_template: "[{timestamp}][{source}] {text}"
\ No newline at end of file
diff --git a/src-python/models/translation/translation_translator.py b/src-python/models/translation/translation_translator.py
index 7efc58d6..d311f88a 100644
--- a/src-python/models/translation/translation_translator.py
+++ b/src-python/models/translation/translation_translator.py
@@ -425,9 +425,18 @@ class Translator:
                 target_language = translation_lang[translator_name]["target"][target_language]
         return source_language, target_language
 
-    def translate(self, translator_name: str, weight_type: str, source_language: str, target_language: str, target_country: str, message: str) -> Any:
+    def translate(self, translator_name: str, weight_type: str, source_language: str, target_language: str, target_country: str, message: str, context_history: Optional[list[dict]] = None) -> Any:
         """Translate `message` using the named translator backend.
 
+        Args:
+            translator_name: Name of the translator backend to use
+            weight_type: Model weight type for CTranslate2
+            source_language: Source language name
+            target_language: Target language name
+            target_country: Target country for locale-specific translations
+            message: Text to translate
+            context_history: Optional conversation context (Chat/Mic/Speaker messages)
+
         Returns translated string on success, or False on failure. When
         source_language == target_language the original message is returned.
         """
@@ -460,6 +469,8 @@ class Translator:
                     if self.plamo_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.plamo_client.setContextHistory(context_history)
                         result = self.plamo_client.translate(
                             message,
                             input_lang=source_language,
@@ -469,6 +480,8 @@ class Translator:
                     if self.gemini_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.gemini_client.setContextHistory(context_history)
                         result = self.gemini_client.translate(
                             message,
                             input_lang=source_language,
@@ -478,6 +491,8 @@ class Translator:
                     if self.openai_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.openai_client.setContextHistory(context_history)
                         result = self.openai_client.translate(
                             message,
                             input_lang=source_language,
@@ -487,6 +502,8 @@ class Translator:
                     if self.groq_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.groq_client.setContextHistory(context_history)
                         result = self.groq_client.translate(
                             message,
                             input_lang=source_language,
@@ -496,6 +513,8 @@ class Translator:
                     if self.openrouter_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.openrouter_client.setContextHistory(context_history)
                         result = self.openrouter_client.translate(
                             message,
                             input_lang=source_language,
@@ -505,6 +524,8 @@ class Translator:
                     if self.lmstudio_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.lmstudio_client.setContextHistory(context_history)
                         result = self.lmstudio_client.translate(
                             message,
                             input_lang=source_language,
@@ -514,6 +535,8 @@ class Translator:
                     if self.ollama_client is None:
                         result = False
                     else:
+                        if context_history:
+                            self.ollama_client.setContextHistory(context_history)
                         result = self.ollama_client.translate(
                             message,
                             input_lang=source_language,
diff --git a/src-python/models/translation/translation_utils.py b/src-python/models/translation/translation_utils.py
index 8c3e4e46..23f24b15 100644
--- a/src-python/models/translation/translation_utils.py
+++ b/src-python/models/translation/translation_utils.py
@@ -101,16 +101,16 @@ def downloadCTranslate2Tokenizer(path: str, weight_type: str = "m2m100_418M-ct2-
         tokenizer_path = os_path.join("./weights", "ctranslate2", directory_name, "tokenizer")
         transformers.AutoTokenizer.from_pretrained(tokenizer, cache_dir=tokenizer_path)
 
-def loadPromptConfig(root_path: str | None = None, prompt_filename: str | None = None) -> dict:
+def loadTranslatePromptConfig(root_path: str | None = None, prompt_filename: str | None = None) -> dict:
     # PyInstaller 展開後
-    if root_path and prompt_filename and os_path.exists(os_path.join(root_path, "_internal", "prompt", prompt_filename)):
-        prompt_path = os_path.join(root_path, "_internal", "prompt", prompt_filename)
+    if root_path and prompt_filename and os_path.exists(os_path.join(root_path, "_internal", "translation_settings", "prompt", prompt_filename)):
+        prompt_path = os_path.join(root_path, "_internal", "translation_settings", "prompt", prompt_filename)
     # src-python 直下実行
-    elif prompt_filename and os_path.exists(os_path.join(os_path.dirname(__file__), "models", "translation", "prompt", prompt_filename)):
-        prompt_path = os_path.join(os_path.dirname(__file__), "models", "translation", "prompt", prompt_filename)
+    elif prompt_filename and os_path.exists(os_path.join(os_path.dirname(__file__), "models", "translation", "translation_settings", "prompt", prompt_filename)):
+        prompt_path = os_path.join(os_path.dirname(__file__), "models", "translation", "translation_settings", "prompt", prompt_filename)
     # translation フォルダ直下実行
-    elif prompt_filename and os_path.exists(os_path.join(os_path.dirname(__file__), "prompt", prompt_filename)):
-        prompt_path = os_path.join(os_path.dirname(__file__), "prompt", prompt_filename)
+    elif prompt_filename and os_path.exists(os_path.join(os_path.dirname(__file__), "translation_settings", "prompt", prompt_filename)):
+        prompt_path = os_path.join(os_path.dirname(__file__), "translation_settings", "prompt", prompt_filename)
     else:
         raise FileNotFoundError(f"Prompt file not found: {prompt_filename}")
     with open(prompt_path, "r", encoding="utf-8") as f: