diff --git a/i18n/en/msg.txt b/i18n/en/msg.txt
new file mode 100644
index 0000000..ab8e5e3
--- /dev/null
+++ b/i18n/en/msg.txt
@@ -0,0 +1,77 @@
+AI 推理控制台: AI Reasoning Console
+API Endpoint (可选): API Endpoint (Optional)
+API Key: API Key
+API endpoint URL (optional, uses default if empty): API endpoint URL (optional, uses default if empty)
+API key for LLM provider: API key for LLM provider
+Cancel: Cancel
+Conform: Confirm
+Creation timestamp: Creation timestamp
+Discard: Discard
+Enable automatic error recovery: Enable automatic error recovery
+Enable automatic session search: Enable automatic session search
+Enable automatic skill loading: Enable automatic skill loading
+Error: Error
+Hermes Reasoning module configuration settings: Hermes Reasoning module configuration settings
+Invalid request: Invalid request
+JSON-encoded detailed reasoning steps: JSON-encoded detailed reasoning steps
+JSON-encoded execution plan: JSON-encoded execution plan
+JSON-encoded final decision metadata: JSON-encoded final decision metadata
+JSON-encoded safety violations detected: JSON-encoded safety violations detected
+LLM model name (e.g. qwen3-max, claude-sonnet-4): LLM model name (e.g. qwen3-max, claude-sonnet-4)
+LLM temperature (0.0-1.0): LLM temperature (0.0-1.0)
+Last update timestamp: Last update timestamp
+Maximum output tokens: Maximum output tokens
+Maximum reasoning steps per task: Maximum reasoning steps per task
+Maximum recovery attempts per error: Maximum recovery attempts per error
+Maximum tokens for reasoning context: Maximum tokens for reasoning context
+Maximum tool calls per reasoning step: Maximum tool calls per reasoning step
+Model provider (e.g. openrouter, anthropic, dashscope): Model provider (e.g. openrouter, anthropic, dashscope)
+Original user request: Original user request
+Reasoning sessions with execution plans and context awareness: Reasoning sessions with execution plans and context awareness
+Reset: Reset
+Safety mode\x3A strict, moderate, lenient: Safety mode\x3A strict, moderate, lenient
+Session status\x3A pending, executing, completed, failed, blocked, cancelled: Session status\x3A pending, executing, completed, failed, blocked, cancelled
+Submit: Submit
+Success: Success
+Summary of gathered context: Summary of gathered context
+System prompt template for reasoning: System prompt template for reasoning
+Temperature (0.0-1.0): Temperature (0.0-1.0)
+Top-p (0.0-1.0): Top-p (0.0-1.0)
+Top-p sampling parameter: Top-p sampling parameter
+Unique configuration identifier: Unique configuration identifier
+Unique reasoning session identifier: Unique reasoning session identifier
+User ID for multi-user isolation: User ID for multi-user isolation
+上下文摘要: Context Summary
+保存失败\x3A : Save Failed\x3A 
+保存引擎配置: Save Engine Configuration
+保存模型配置: Save Model Configuration
+创建时间: Creation Time
+启用技能自动加载: Enable Automatic Skill Loading
+启用跨会话搜索: Enable Cross-Session Search
+启用错误恢复: Enable Error Recovery
+安全模式: Safety Mode
+开始推理: Start Reasoning
+推理交互: Reasoning Interaction
+推理会话: Reasoning Session
+推理失败: Reasoning Failed
+推理完成: Reasoning Completed
+推理异常\x3A : Reasoning Exception\x3A 
+推理引擎: Reasoning Engine
+推理控制台: Reasoning Console
+推理请求失败: Reasoning Request Failed
+推理配置: Reasoning Configuration
+最大上下文 token 数: Maximum Context Tokens
+最大恢复尝试次数: Maximum Recovery Attempts
+最大推理步骤数: Maximum Reasoning Steps
+最大输出 token 数: Maximum Output Tokens
+模型名称: Model Name
+模型提供商: Model Provider
+模型配置: Model Configuration
+每步最大工具调用数: Maximum Tool Calls Per Step
+清空日志: Clear Logs
+状态: Status
+系统提示词模板: System Prompt Template
+请求内容: Request Content
+请输入推理请求: Please Enter Reasoning Request
+配置保存成功: Configuration Saved Successfully
+配置管理: Configuration Management
diff --git a/i18n/jp/msg.txt b/i18n/jp/msg.txt
new file mode 100644
index 0000000..370ce50
--- /dev/null
+++ b/i18n/jp/msg.txt
@@ -0,0 +1,77 @@
+AI 推理控制台: AI推論コンソール
+API Endpoint (可选): APIエンドポイント（オプション）
+API Key: APIキー
+API endpoint URL (optional, uses default if empty): APIエンドポイントURL（オプション、空の場合はデフォルトを使用）
+API key for LLM provider: LLMプロバイダーのAPIキー
+Cancel: キャンセル
+Conform: 確認
+Creation timestamp: 作成タイムスタンプ
+Discard: 破棄
+Enable automatic error recovery: 自動エラー回復を有効にする
+Enable automatic session search: 自動セッション検索を有効にする
+Enable automatic skill loading: 自動スキル読み込みを有効にする
+Error: エラー
+Hermes Reasoning module configuration settings: Hermes推論モジュール設定
+Invalid request: 無効なリクエスト
+JSON-encoded detailed reasoning steps: JSONエンコードされた詳細な推論ステップ
+JSON-encoded execution plan: JSONエンコードされた実行計画
+JSON-encoded final decision metadata: JSONエンコードされた最終決定メタデータ
+JSON-encoded safety violations detected: JSONエンコードされた検出された安全違反
+LLM model name (e.g. qwen3-max, claude-sonnet-4): LLMモデル名（例：qwen3-max、claude-sonnet-4）
+LLM temperature (0.0-1.0): LLM温度（0.0-1.0）
+Last update timestamp: 最終更新タイムスタンプ
+Maximum output tokens: 最大出力トークン
+Maximum reasoning steps per task: タスクあたりの最大推論ステップ数
+Maximum recovery attempts per error: エラーあたりの最大回復試行回数
+Maximum tokens for reasoning context: 推論コンテキストの最大トークン数
+Maximum tool calls per reasoning step: 推論ステップあたりの最大ツール呼び出し数
+Model provider (e.g. openrouter, anthropic, dashscope): モデルプロバイダー（例：openrouter、anthropic、dashscope）
+Original user request: 元のユーザーリクエスト
+Reasoning sessions with execution plans and context awareness: 実行計画とコンテキスト認識を持つ推論セッション
+Reset: リセット
+Safety mode\x3A strict, moderate, lenient: セーフティモード\x3A 厳格、中程度、寛容
+Session status\x3A pending, executing, completed, failed, blocked, cancelled: セッションステータス\x3A 保留中、実行中、完了、失敗、ブロック、キャンセル
+Submit: 送信
+Success: 成功
+Summary of gathered context: 収集されたコンテキストの要約
+System prompt template for reasoning: 推論用システムプロンプトテンプレート
+Temperature (0.0-1.0): 温度（0.0-1.0）
+Top-p (0.0-1.0): Top-p（0.0-1.0）
+Top-p sampling parameter: Top-pサンプリングパラメータ
+Unique configuration identifier: 一意な設定識別子
+Unique reasoning session identifier: 一意な推論セッション識別子
+User ID for multi-user isolation: マルチユーザー分離用ユーザーID
+上下文摘要: コンテキスト要約
+保存失败\x3A : 保存失敗\x3A 
+保存引擎配置: エンジン設定を保存
+保存模型配置: モデル設定を保存
+创建时间: 作成時間
+启用技能自动加载: スキル自動読み込みを有効にする
+启用跨会话搜索: クロスセッション検索を有効にする
+启用错误恢复: エラー回復を有効にする
+安全模式: セーフティモード
+开始推理: 推論開始
+推理交互: 推論インタラクション
+推理会话: 推論セッション
+推理失败: 推論失敗
+推理完成: 推論完了
+推理异常\x3A : 推論例外\x3A 
+推理引擎: 推論エンジン
+推理控制台: 推論コンソール
+推理请求失败: 推論リクエスト失敗
+推理配置: 推論設定
+最大上下文 token 数: 最大コンテキストトークン数
+最大恢复尝试次数: 最大回復試行回数
+最大推理步骤数: 最大推論ステップ数
+最大输出 token 数: 最大出力トークン数
+模型名称: モデル名
+模型提供商: モデルプロバイダー
+模型配置: モデル設定
+每步最大工具调用数: ステップあたりの最大ツール呼び出し数
+清空日志: ログをクリア
+状态: ステータス
+系统提示词模板: システムプロンプトテンプレート
+请求内容: リクエスト内容
+请输入推理请求: 推論リクエストを入力してください
+配置保存成功: 設定が正常に保存されました
+配置管理: 設定管理
diff --git a/i18n/ko/msg.txt b/i18n/ko/msg.txt
new file mode 100644
index 0000000..84dc8b5
--- /dev/null
+++ b/i18n/ko/msg.txt
@@ -0,0 +1,77 @@
+AI 推理控制台: AI 추론 콘솔
+API Endpoint (可选): API 엔드포인트 (선택)
+API Key: API 키
+API endpoint URL (optional, uses default if empty): API 엔드포인트 URL (선택, 비어 있으면 기본값 사용)
+API key for LLM provider: LLM 제공업체용 API 키
+Cancel: 취소
+Conform: 확인
+Creation timestamp: 생성 타임스탬프
+Discard: 폐기
+Enable automatic error recovery: 자동 오류 복구 활성화
+Enable automatic session search: 자동 세션 검색 활성화
+Enable automatic skill loading: 자동 스킬 로드 활성화
+Error: 오류
+Hermes Reasoning module configuration settings: Hermes 추론 모듈 구성 설정
+Invalid request: 잘못된 요청
+JSON-encoded detailed reasoning steps: JSON 인코딩된 상세 추론 단계
+JSON-encoded execution plan: JSON 인코딩된 실행 계획
+JSON-encoded final decision metadata: JSON 인코딩된 최종 결정 메타데이터
+JSON-encoded safety violations detected: JSON 인코딩된 감지된 안전 위반
+LLM model name (e.g. qwen3-max, claude-sonnet-4): LLM 모델 이름 (예: qwen3-max, claude-sonnet-4)
+LLM temperature (0.0-1.0): LLM 온도 (0.0-1.0)
+Last update timestamp: 최종 업데이트 타임스탬프
+Maximum output tokens: 최대 출력 토큰
+Maximum reasoning steps per task: 작업당 최대 추론 단계 수
+Maximum recovery attempts per error: 오류당 최대 복구 시도 횟수
+Maximum tokens for reasoning context: 추론 컨텍스트 최대 토큰 수
+Maximum tool calls per reasoning step: 추론 단계당 최대 도구 호출 수
+Model provider (e.g. openrouter, anthropic, dashscope): 모델 제공업체 (예: openrouter, anthropic, dashscope)
+Original user request: 원본 사용자 요청
+Reasoning sessions with execution plans and context awareness: 실행 계획 및 컨텍스트 인식을 갖춘 추론 세션
+Reset: 초기화
+Safety mode\x3A strict, moderate, lenient: 안전 모드\x3A 엄격, 보통, 관대
+Session status\x3A pending, executing, completed, failed, blocked, cancelled: 세션 상태\x3A 대기 중, 실행 중, 완료, 실패, 차단, 취소
+Submit: 제출
+Success: 성공
+Summary of gathered context: 수집된 컨텍스트 요약
+System prompt template for reasoning: 추론용 시스템 프롬프트 템플릿
+Temperature (0.0-1.0): 온도 (0.0-1.0)
+Top-p (0.0-1.0): Top-p (0.0-1.0)
+Top-p sampling parameter: Top-p 샘플링 매개변수
+Unique configuration identifier: 고유 구성 식별자
+Unique reasoning session identifier: 고유 추론 세션 식별자
+User ID for multi-user isolation: 다중 사용자 격리용 사용자 ID
+上下文摘要: 컨텍스트 요약
+保存失败\x3A : 저장 실패\x3A 
+保存引擎配置: 엔진 구성 저장
+保存模型配置: 모델 구성 저장
+创建时间: 생성 시간
+启用技能自动加载: 스킬 자동 로드 활성화
+启用跨会话搜索: 크로스 세션 검색 활성화
+启用错误恢复: 오류 복구 활성화
+安全模式: 안전 모드
+开始推理: 추론 시작
+推理交互: 추론 상호작용
+推理会话: 추론 세션
+推理失败: 추론 실패
+推理完成: 추론 완료
+推理异常\x3A : 추론 예외\x3A 
+推理引擎: 추론 엔진
+推理控制台: 추론 콘솔
+推理请求失败: 추론 요청 실패
+推理配置: 추론 구성
+最大上下文 token 数: 최대 컨텍스트 토큰 수
+最大恢复尝试次数: 최대 복구 시도 횟수
+最大推理步骤数: 최대 추론 단계 수
+最大输出 token 数: 최대 출력 토큰 수
+模型名称: 모델 이름
+模型提供商: 모델 제공업체
+模型配置: 모델 구성
+每步最大工具调用数: 단계당 최대 도구 호출 수
+清空日志: 로그 지우기
+状态: 상태
+系统提示词模板: 시스템 프롬프트 템플릿
+请求内容: 요청 내용
+请输入推理请求: 추론 요청을 입력하세요
+配置保存成功: 구성이 성공적으로 저장되었습니다
+配置管理: 구성 관리
diff --git a/i18n/zh/msg.txt b/i18n/zh/msg.txt
new file mode 100644
index 0000000..8cdc4ae
--- /dev/null
+++ b/i18n/zh/msg.txt
@@ -0,0 +1,77 @@
+AI 推理控制台: AI 推理控制台
+API Endpoint (可选): API Endpoint (可选)
+API Key: API Key
+API endpoint URL (optional, uses default if empty): API endpoint URL (optional, uses default if empty)
+API key for LLM provider: API key for LLM provider
+Cancel: Cancel
+Conform: Conform
+Creation timestamp: Creation timestamp
+Discard: Discard
+Enable automatic error recovery: Enable automatic error recovery
+Enable automatic session search: Enable automatic session search
+Enable automatic skill loading: Enable automatic skill loading
+Error: Error
+Hermes Reasoning module configuration settings: Hermes Reasoning module configuration settings
+Invalid request: Invalid request
+JSON-encoded detailed reasoning steps: JSON-encoded detailed reasoning steps
+JSON-encoded execution plan: JSON-encoded execution plan
+JSON-encoded final decision metadata: JSON-encoded final decision metadata
+JSON-encoded safety violations detected: JSON-encoded safety violations detected
+LLM model name (e.g. qwen3-max, claude-sonnet-4): LLM model name (e.g. qwen3-max, claude-sonnet-4)
+LLM temperature (0.0-1.0): LLM temperature (0.0-1.0)
+Last update timestamp: Last update timestamp
+Maximum output tokens: Maximum output tokens
+Maximum reasoning steps per task: Maximum reasoning steps per task
+Maximum recovery attempts per error: Maximum recovery attempts per error
+Maximum tokens for reasoning context: Maximum tokens for reasoning context
+Maximum tool calls per reasoning step: Maximum tool calls per reasoning step
+Model provider (e.g. openrouter, anthropic, dashscope): Model provider (e.g. openrouter, anthropic, dashscope)
+Original user request: Original user request
+Reasoning sessions with execution plans and context awareness: Reasoning sessions with execution plans and context awareness
+Reset: Reset
+Safety mode\x3A strict, moderate, lenient: Safety mode\x3A strict, moderate, lenient
+Session status\x3A pending, executing, completed, failed, blocked, cancelled: Session status\x3A pending, executing, completed, failed, blocked, cancelled
+Submit: Submit
+Success: Success
+Summary of gathered context: Summary of gathered context
+System prompt template for reasoning: System prompt template for reasoning
+Temperature (0.0-1.0): Temperature (0.0-1.0)
+Top-p (0.0-1.0): Top-p (0.0-1.0)
+Top-p sampling parameter: Top-p sampling parameter
+Unique configuration identifier: Unique configuration identifier
+Unique reasoning session identifier: Unique reasoning session identifier
+User ID for multi-user isolation: User ID for multi-user isolation
+上下文摘要: 上下文摘要
+保存失败\x3A : 保存失败\x3A 
+保存引擎配置: 保存引擎配置
+保存模型配置: 保存模型配置
+创建时间: 创建时间
+启用技能自动加载: 启用技能自动加载
+启用跨会话搜索: 启用跨会话搜索
+启用错误恢复: 启用错误恢复
+安全模式: 安全模式
+开始推理: 开始推理
+推理交互: 推理交互
+推理会话: 推理会话
+推理失败: 推理失败
+推理完成: 推理完成
+推理异常\x3A : 推理异常\x3A 
+推理引擎: 推理引擎
+推理控制台: 推理控制台
+推理请求失败: 推理请求失败
+推理配置: 推理配置
+最大上下文 token 数: 最大上下文 token 数
+最大恢复尝试次数: 最大恢复尝试次数
+最大推理步骤数: 最大推理步骤数
+最大输出 token 数: 最大输出 token 数
+模型名称: 模型名称
+模型提供商: 模型提供商
+模型配置: 模型配置
+每步最大工具调用数: 每步最大工具调用数
+清空日志: 清空日志
+状态: 状态
+系统提示词模板: 系统提示词模板
+请求内容: 请求内容
+请输入推理请求: 请输入推理请求
+配置保存成功: 配置保存成功
+配置管理: 配置管理