xiewoc · xiewoc · Nov 9, 2025 · Nov 9, 2025
diff --git a/README.md b/README.md
@@ -27,32 +27,22 @@
 - 调用本地运行的 Ollama 服务（或其他 AI 接口）
 - 生成回复后通过同一网络回传
 
-### 启动方式
-
-1. **启动 Ollama 服务**：
-   ```bash
-   ollama serve
-   ```
-   > 这会启动 Ollama 后台服务，默认监听 `11434` 端口。
-
-2. （可选）提前下载模型：
-   ```bash
-   ollama pull qwen2.5:7b
-   ```
-   或使用其他轻量模型如 `phi3`、`tinyllama`。
-
-3. 运行 AI 节点程序：
-   ```bash
-   python main.py
-   ```
-
-> 注：Ollama 在首次请求时会自动下载并加载模型（如果未提前 pull）。确保设备有足够存储和内存。
+## 技术规格
+| 连接方式 | 串口 |
+| ----- | -----|
+| 记忆 | 未实现|
+| LLM Tool | 未实现 |
+| 语言支持 | 中、英、日、法、俄、韩、西、德|
+| 服务提供商 | OpenAI（及类似，如：DeepSeek、Ollama）、web sockets、fastapi|
 
 ### 当前配置示例
 
 ```json
 {
   "platform": "ollama",
+  "localization":{
+      "language": "zh_CN"
+  },
   "api_keys": {
     "openai": "your-openai-api-key",
     "deepseek": "your-deepseek-api-key",
@@ -84,7 +74,7 @@
 >
 >如果你在使用OpenRouter，请参照[README_OPENROUTER](README_OPENROUTER.md)
 >
->若要接入 `AstrBot` ，可以使用 [AstrBot适配器](https://github.com/xiewoc/astrbot_plugin_adapter_meshbot)
+>若要接入 `AstrBot` ，可以使用 [AstrBot适配器](https://github.com/xiewoc/astrbot_plugin_adapter_meshbot) (*推荐*)
 
 完全可以在树莓派 + TTGO T-Beam 上跑起来，边走边聊。
 
@@ -100,6 +90,11 @@
    ```bash
    ollama serve
    ```
+
+   > 这会启动 Ollama 后台服务，默认监听 `11434` 端口。
+   >
+   > 注：Ollama 在首次请求时会自动下载并加载模型（如果未提前 pull）。确保设备有足够存储和内存。
+
 5. 运行主程序：
    ```bash
    python main.py
@@ -111,16 +106,16 @@
 
 ## 🎈当前版本 
 
-V 1.0.3
+V 1.0.3 - pre 1
 
-- 重构了文件夹结构
-- 添加了`Gemini`, `SiliconFlow`, `Claude`和`Fastapi`的适配器
-- 重构了`config.json`
+- 添加了`localization` ,支持了 `en`, `zh_CN`, `ru`, `jp`, `ko`, `es`, `de`
 
 ## 🌱 后续想法
 
 - 引入上下文记忆，让对话更连贯
 - 添加一个WebUI
+- 添加LLM Tool
+- 优化`api`文件夹
 
 ## 🙏 写在最后
 

diff --git a/README_EN.md b/README_EN.md
@@ -1,58 +1,48 @@
 <div align="center">
 
-[**简体中文**](README.md) | **English**
+**English** | [ **简体中文** ](README.md)
 
 </div>
 
 # Mesh AI Assistant
 
-A small AI node that quietly resides in the Mesh network.  
-You send it a message, and it replies with a sentence.
+A small AI node quietly residing in the Mesh network.  
+You send it a message, and it replies.
 
 Unobtrusive, offline, and serverless.  
-Just for those times when you're in the mountains, in the wild, or somewhere with no signal, and you can still ask, "What do you think?" and receive an answer.
+Just for those times when you're in the mountains, the wilderness, or places with no signal – you can still ask, "What do you think?" and receive an answer.
 
 ## 🧩 What Can It Do?
 
--   Receive private messages sent to it (peer-to-peer messages)
--   Generate short replies using a local AI model
--   Send the response back the same way, as if it's always online waiting for you
+- Receive private messages sent to it (peer-to-peer)
+- Generate short replies using a local AI model
+- Send the response back through the same path, as if it's always online waiting for you
 
-All processing is done locally, ensuring privacy and control.
+All processing is done locally, ensuring privacy control.
 
 ## ⚙️ Technical Implementation
 
--   Uses Python to listen for serial port messages from Meshtastic devices
--   Extracts content when a private message for this node is received
--   Calls a locally running Ollama service (or other AI interfaces)
--   Sends the generated reply back through the same network
+- Uses Python to listen for serial port messages from Meshtastic devices
+- Extracts content when a private message for this node is received
+- Calls a locally running Ollama service (or other AI interface)
+- Sends the generated reply back through the same network
 
-### How to Start
-
-1.  **Start the Ollama Service**:
-    ```bash
-    ollama serve
-    ```
-    > This starts the Ollama background service, listening on port `11434` by default.
-
-2.  (Optional) Download a model in advance:
-    ```bash
-    ollama pull qwen2.5:7b
-    ```
-    Or use other lightweight models like `phi3` or `tinyllama`.
-
-3.  Run the AI node program:
-    ```bash
-    python main.py
-    ```
-
-> Note: Ollama will automatically download and load the model on the first request (if not pulled in advance). Ensure your device has sufficient storage and memory.
+## Technical Specifications
+| Connection | Serial Port |
+| ----- | -----|
+| Memory | Not Implemented |
+| LLM Tools | Not Implemented |
+| Language Support | Chinese, English, Japanese, French, Russian, Korean, Spanish, German |
+| Service Providers | OpenAI (and similar, e.g., DeepSeek, Ollama), web sockets, FastAPI |
 
 ### Current Configuration Example
 
 ```json
 {
   "platform": "ollama",
+  "localization":{
+      "language": "en"
+  },
   "api_keys": {
     "openai": "your-openai-api-key",
     "deepseek": "your-deepseek-api-key",
@@ -79,12 +69,12 @@ All processing is done locally, ensuring privacy and control.
 }
 ```
 
-> [!IMPORTANT]
-> Please replace `your-api-key` with your actual API key when using services like `openai`, `deepseek`, etc.
+>[!IMPORTANT]
+>Please replace `your-api-key` with your actual API key when using services like `openai`, `deepseek`, etc.
 >
-> If you are using OpenRouter, please refer to [README_OPENROUTER](README_OPENROUTER.md)
+>If you are using OpenRouter, please refer to [README_OPENROUTER](README_OPENROUTER.md)
 >
-> To integrate with `AstrBot`, you can use the [AstrBot Adapter](https://github.com/xiewoc/astrbot_plugin_adapter_meshbot)
+>To integrate with `AstrBot`, you can use the [AstrBot Adapter](https://github.com/xiewoc/astrbot_plugin_adapter_meshbot) (*Recommended*)
 
 It can easily run on a Raspberry Pi + TTGO T-Beam, allowing you to chat on the go.
 
@@ -100,37 +90,42 @@ It can easily run on a Raspberry Pi + TTGO T-Beam, allowing you to chat on the g
     ```bash
     ollama serve
     ```
+
+    > This starts the Ollama background service, listening on port `11434` by default.
+    >
+    > Note: Ollama will automatically download and load the model on the first request (if not pulled beforehand). Ensure the device has sufficient storage and memory.
+
 5.  Run the main program:
     ```bash
     python main.py
     ```
-6.  Send a private message to it from another device and wait for a reply.
+6.  Send a private message to it from another device and wait for the reply.
 
-> [!IMPORTANT]
-> Please pay attention to the runtime path when executing the main program; it must be run from within the project folder.
+>[!IMPORTANT]
+>Please pay attention to the working directory when running the main program; it must be within the project folder.
 
 ## 🎈 Current Version
 
-V 1.0.3
+V 1.0.3 - pre 1
 
--   Refactored the folder structure
--   Added adapters for `Gemini`, `SiliconFlow`, `Claude`, and `Fastapi`
--   Refactored `config.json`
+- Added `localization`, supporting `en`, `zh_CN`, `ru`, `jp`, `ko`, `es`, `de`
 
 ## 🌱 Future Ideas
 
--   Introduce context memory for more coherent conversations
--   Add a WebUI
+- Introduce context memory for more coherent conversations
+- Add a WebUI
+- Add LLM Tools
+- Optimize the `api` folder
 
 ## 🙏 Final Words
 
 This project isn't meant to replace anyone, nor is it about creating an overly intelligent AI.  
-It's just about leaving a voice that can respond to you in those quiet places.
+It's simply about leaving a voice that can respond to you in those quiet places.
 
 If you also appreciate this concept, you're welcome to help improve it.
 
-Simultaneously, thanks to the developers who have contributed to this project; we appreciate your support and efforts.
+Special thanks to the developers who have contributed to this project – we appreciate your support and dedication.
 
-May your Meshtastic node run stably in the mountains and wilds, where every reply is like a quietly lit signal lamp. 📡💡
+May your Meshtastic node run stably out in the wilds, where every reply is like a quietly lit signal lamp. 📡💡
 
 Happy Exploring! ✨
diff --git a/config.json b/config.json
@@ -1,5 +1,8 @@
 {
   "platform": "ollama",
+  "localization":{
+      "language": "zh_CN"
+  },
   "api_keys": {
     "openai": "your-openai-api-key",
     "deepseek": "your-deepseek-api-key",

diff --git a/main.py b/main.py
@@ -4,9 +4,11 @@
 import os
 from pathlib import Path
 
-from meshbot.core.bot import MeshAIBot
+from meshbot.core.bot import MeshBot
 from meshbot.handlers.signal_handlers import setup_signal_handlers
+from meshbot.config.config_loader import load_config
 from meshbot.config.config_loader import create_example_config
+from meshbot.utils.localize import i18n
 
 # 日志配置
 logging.basicConfig(
@@ -15,7 +17,6 @@
 )
 logger = logging.getLogger(__name__)
 
-
 def check_config():
     """检查配置文件是否存在，如果不存在则创建示例"""
     config_path = Path(__file__).parent / "config.json"
@@ -33,12 +34,13 @@ async def main() -> None:
     if not check_config():
         return
 
-    bot = MeshAIBot()
+    load_config()
+    bot = MeshBot()
     setup_signal_handlers(bot)
     try:
         await bot.run()
     except Exception as e:
-        logger.error(f"💥 机器人运行异常: {e}")
+        logger.error(i18n.gettext('bot_running_error',err = e))
     finally:
         await bot.shutdown()