diff --git a/README.md b/README.md
index f4e17df4..abe012e5 100644
--- a/README.md
+++ b/README.md
@@ -232,6 +232,48 @@ python scripts/check_deployment_cn.py --base-url http://你的IP:你的端口/v1
 
 ## 使用 AutoGLM
 
+### Web UI 界面
+
+Phone Agent 提供了可视化的 Web 界面，无需命令行即可轻松使用。
+
+#### 启动 WebUI
+
+```bash
+python web_ui.py
+```
+
+启动后，在浏览器中访问 [http://127.0.0.1:7860](http://127.0.0.1:7860)
+
+![WebUI界面](resources/iShot_2025-12-14_16.27.01.png)
+
+#### 📖 使用指南
+
+**快速开始**
+
+1. **检查环境**: 前往"系统检查"标签页，运行系统检查确保环境配置正确
+2. **配置设置**: 在"配置"区域设置模型API地址和API Key
+3. **选择设备**: 点击"刷新设备"按钮，从下拉列表中选择目标设备
+4. **输入任务**: 在任务输入框中用自然语言描述您的需求
+5. **执行任务**: 点击"执行任务"按钮，系统会自动控制手机完成任务
+
+**功能说明**
+
+💬 **对话控制**
+- 任务执行: 支持流式输出，实时显示AI的思考过程和执行动作
+- 截图显示: 右侧实时显示设备当前屏幕状态
+- 新对话: 清空当前上下文，开始全新的任务会话
+
+🔍 **系统检查**
+- 检查ADB工具安装状态
+- 检查设备连接状态
+- 检查ADB Keyboard安装状态
+- 检查模型API连接状态
+
+📱 **设备管理**
+- 查看所有已连接设备
+- 支持USB和WiFi连接
+- 远程设备连接功能
+
 ### 命令行
 
 根据你部署的模型, 设置 `--base-url` 和 `--model` 参数. 例如:
diff --git a/README_en.md b/README_en.md
index e2e78203..4c3fd433 100644
--- a/README_en.md
+++ b/README_en.md
@@ -182,6 +182,62 @@ python scripts/check_deployment_en.py --base-url https://api.parasail.io/v1 --mo
 Upon successful execution, the script will display the model's inference result and token statistics, helping you confirm whether the model deployment is working correctly.
 
 ## Using AutoGLM
+### Web UI Interface
+
+Phone Agent provides a visual web interface that can be easily used without the need for command lines.
+
+#### Starting WebUI
+
+```bash
+python web_ui.py
+
+```
+
+After starting, access [http://127.0.0.1:7860](http://127.0.0.1:7860) in your browser
+
+![WebUI Interface](resources/iShot_2025-12-14_16.27.01.png)
+
+#### 📖 User Guide
+
+**Quick Start**
+
+1. **Check Environment**: Go to the "System Check" tab and run a system check to ensure the environment is configured correctly.
+
+2. **Configuration Settings**: Set the model API address and API Key in the "Configuration" area.
+
+3. **Select Device**: Click the "Refresh Device" button and select the target device from the drop-down list.
+
+4. **Enter Task**: Describe your requirements in natural language in the task input box.
+
+5. **Execute Task**: Click the "Execute Task" button, and the system will automatically control the phone to complete the task.
+
+**Function Description**
+
+💬 **Dialogue Control**
+
+- Task Execution: Supports streaming output, displaying the AI's thought process and actions in real time.
+
+- Screenshot Display: Real-time display of the device's current screen status on the right.
+
+- New Dialogue: Clears the current context and starts a completely new task session.
+
+🔍 **System Checks**
+
+- Check ADB tool installation status
+
+- Check device connection status
+
+- Check ADB Keyboard installation status
+
+- Check model API connection status
+
+📱 **Device Management**
+
+- View all connected devices
+
+- Supports USB and WiFi connections
+
+- Remote device connection function
 
 ### Command Line
 
diff --git a/requirements.txt b/requirements.txt
index 16aa4265..6f062eac 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,5 +1,6 @@
 Pillow>=12.0.0
 openai>=2.9.0
+gradio>=5.18.0
 
 # For Model Deployment
 
diff --git a/resources/iShot_2025-12-14_16.27.01.png b/resources/iShot_2025-12-14_16.27.01.png
new file mode 100644
index 00000000..011db080
Binary files /dev/null and b/resources/iShot_2025-12-14_16.27.01.png differ
diff --git a/web_ui.py b/web_ui.py
new file mode 100644
index 00000000..014eec85
--- /dev/null
+++ b/web_ui.py
@@ -0,0 +1,887 @@
+#!/usr/bin/env python3
+"""
+Gradio Web UI for Phone Agent - AI-powered phone automation.
+
+This provides a user-friendly web interface for controlling and monitoring
+the phone agent assistant.
+"""
+
+import base64
+import json
+import os
+import shutil
+import subprocess
+import tempfile
+import time
+import traceback
+from datetime import datetime
+from typing import Generator, List, Tuple
+
+import gradio as gr
+from openai import OpenAI
+
+from phone_agent import PhoneAgent
+from phone_agent.adb import get_screenshot, list_devices
+from phone_agent.agent import AgentConfig, StepResult
+from phone_agent.model import ModelConfig
+
+
+# ============================================================================
+# System Check Functions
+# ============================================================================
+
+def check_adb_installation() -> Tuple[bool, str]:
+    """Check if ADB is installed and accessible."""
+    if shutil.which("adb") is None:
+        return False, "❌ ADB未安装或不在PATH中"
+
+    try:
+        result = subprocess.run(
+            ["adb", "version"], capture_output=True, text=True, timeout=10
+        )
+        if result.returncode == 0:
+            version_line = result.stdout.strip().split("\n")[0]
+            return True, f"✅ {version_line}"
+        else:
+            return False, "❌ ADB命令运行失败"
+    except Exception as e:
+        return False, f"❌ ADB检查失败: {str(e)}"
+
+
+def check_device_connection() -> Tuple[bool, str, List[str]]:
+    """Check connected devices."""
+    try:
+        devices = list_devices()
+        if not devices:
+            return False, "❌ 没有连接的设备", []
+
+        device_list = [f"{d.device_id} ({d.model or 'Unknown'})" for d in devices]
+        device_ids = [d.device_id for d in devices]
+        return True, f"✅ 找到 {len(devices)} 个设备", device_ids
+    except Exception as e:
+        return False, f"❌ 设备检查失败: {str(e)}", []
+
+
+def check_adb_keyboard(device_id: str = None) -> Tuple[bool, str]:
+    """Check if ADB Keyboard is installed on the device."""
+    try:
+        cmd = ["adb"]
+        if device_id:
+            cmd.extend(["-s", device_id])
+        cmd.extend(["shell", "ime", "list", "-s"])
+
+        result = subprocess.run(cmd, capture_output=True, text=True, timeout=10)
+        ime_list = result.stdout.strip()
+
+        if "com.android.adbkeyboard/.AdbIME" in ime_list:
+            return True, "✅ ADB Keyboard已安装"
+        else:
+            return False, "❌ ADB Keyboard未安装"
+    except Exception as e:
+        return False, f"❌ ADB Keyboard检查失败: {str(e)}"
+
+
+def check_model_api(base_url: str, model_name: str, api_key: str = "EMPTY") -> Tuple[bool, str]:
+    """Check if the model API is accessible."""
+    try:
+        client = OpenAI(base_url=base_url, api_key=api_key, timeout=10.0)
+        models_response = client.models.list()
+        available_models = [model.id for model in models_response.data]
+        return True, f"✅ API连接成功 ({len(available_models)} 个可用模型)"
+    except Exception as e:
+        error_msg = str(e)
+        if "Connection refused" in error_msg or "Connection error" in error_msg:
+            return False, f"❌ 无法连接到 {base_url}"
+        elif "timed out" in error_msg.lower() or "timeout" in error_msg.lower():
+            return False, f"❌ 连接超时: {base_url}"
+        else:
+            return False, f"❌ API错误: {error_msg}"
+
+
+def run_full_system_check(base_url: str, model_name: str, api_key: str, device_id: str = None) -> str:
+    """Run comprehensive system check."""
+    report = ["# 系统环境检查报告", ""]
+    report.append(f"**检查时间**: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
+    report.append("")
+
+    # Check 1: ADB Installation
+    report.append("## 1. ADB安装检查")
+    adb_ok, adb_msg = check_adb_installation()
+    report.append(adb_msg)
+    report.append("")
+
+    if not adb_ok:
+        report.append("**提示**: 请先安装ADB工具")
+        report.append("- macOS: `brew install android-platform-tools`")
+        report.append("- Linux: `sudo apt install android-tools-adb`")
+        report.append("- Windows: 从 [官网](https://developer.android.com/studio/releases/platform-tools) 下载")
+        return "\n".join(report)
+
+    # Check 2: Device Connection
+    report.append("## 2. 设备连接检查")
+    dev_ok, dev_msg, device_list = check_device_connection()
+    report.append(dev_msg)
+    if dev_ok:
+        for dev in device_list:
+            report.append(f"  - {dev}")
+    report.append("")
+
+    if not dev_ok:
+        report.append("**提示**: 请连接Android设备并启用USB调试")
+        return "\n".join(report)
+
+    # Check 3: ADB Keyboard
+    report.append("## 3. ADB Keyboard检查")
+    kbd_ok, kbd_msg = check_adb_keyboard(device_id)
+    report.append(kbd_msg)
+    report.append("")
+
+    if not kbd_ok:
+        report.append("**提示**: 请在设备上安装ADB Keyboard")
+        report.append("  1. 下载: [ADBKeyboard.apk](https://github.com/senzhk/ADBKeyBoard/blob/master/ADBKeyboard.apk)")
+        report.append("  2. 安装: `adb install ADBKeyboard.apk`")
+        report.append("  3. 在设置中启用该输入法")
+        report.append("")
+
+    # Check 4: Model API
+    report.append("## 4. 模型API检查")
+    api_ok, api_msg = check_model_api(base_url, model_name, api_key)
+    report.append(api_msg)
+    report.append("")
+
+    if not api_ok:
+        report.append("**提示**: 请检查模型服务器是否运行，并确认URL和API Key正确")
+        report.append("")
+
+    # Summary
+    report.append("## 检查总结")
+    all_ok = adb_ok and dev_ok and kbd_ok and api_ok
+    if all_ok:
+        report.append("✅ **所有检查通过！可以开始使用Phone Agent**")
+    else:
+        report.append("❌ **部分检查未通过，请根据上述提示修复问题**")
+
+    return "\n".join(report)
+
+
+# ============================================================================
+# Agent Management
+# ============================================================================
+
+class AgentSession:
+    """Manages a phone agent session."""
+
+    def __init__(self, base_url: str, model_name: str, api_key: str,
+                 max_steps: int, device_id: str, lang: str):
+        self.model_config = ModelConfig(
+            base_url=base_url,
+            model_name=model_name,
+            api_key=api_key,
+        )
+
+        self.agent_config = AgentConfig(
+            max_steps=max_steps,
+            device_id=device_id if device_id else None,
+            lang=lang,
+            verbose=True,
+        )
+
+        self.agent = PhoneAgent(
+            model_config=self.model_config,
+            agent_config=self.agent_config,
+        )
+
+        self.current_task = None
+        self.should_stop = False
+
+    def stop_task(self):
+        """Stop the current running task."""
+        self.should_stop = True
+
+
+    def run_task_stream(self, task: str) -> Generator[Tuple[str, str], None, None]:
+        """Run a task and yield updates for streaming display."""
+        self.current_task = task
+        self.should_stop = False
+        self.agent.reset()
+
+        # Yield initial message
+        initial_output = self._format_step_output("开始执行任务", task, "info")
+        yield initial_output, None
+
+        # First step
+        try:
+            result = self.agent.step(task)
+            step_output = self._format_step_result(result, 1)
+            screenshot_path = self._get_screenshot()
+            yield step_output, screenshot_path
+
+            if result.finished:
+                final_msg = result.message or "任务完成"
+                final_output = self._format_step_output("任务完成", final_msg, "success")
+                yield final_output, screenshot_path
+                return
+        except Exception as e:
+            error_msg = f"错误: {str(e)}\n{traceback.format_exc()}"
+            error_output = self._format_step_output("执行错误", error_msg, "error")
+            yield error_output, None
+            return
+
+        # Continue steps
+        while self.agent.step_count < self.agent_config.max_steps:
+            # Check if task should be stopped
+            if self.should_stop:
+                screenshot_path = self._get_screenshot()
+                stop_output = self._format_step_output("任务终止", "任务已被用户手动终止", "warning")
+                yield stop_output, screenshot_path
+                return
+
+            try:
+                result = self.agent.step()
+                step_output = self._format_step_result(result, self.agent.step_count)
+                screenshot_path = self._get_screenshot()
+                yield step_output, screenshot_path
+
+                if result.finished:
+                    final_msg = result.message or "任务完成"
+                    final_output = self._format_step_output("任务完成", final_msg, "success")
+                    yield final_output, screenshot_path
+                    return
+
+            except Exception as e:
+                error_msg = f"错误: {str(e)}\n{traceback.format_exc()}"
+                error_output = self._format_step_output("执行错误", error_msg, "error")
+                yield error_output, None
+                return
+
+        # Max steps reached
+        screenshot_path = self._get_screenshot()
+        warning_output = self._format_step_output("任务终止", "达到最大步数限制", "warning")
+        yield warning_output, screenshot_path
+
+    def _format_step_result(self, result: StepResult, step_num: int) -> str:
+        """Format a step result for display."""
+        lines = []
+        lines.append(f"\n{'='*60}")
+        lines.append(f"📍 **步骤 {step_num}**")
+        lines.append(f"{'='*60}\n")
+
+        lines.append(f"### 💭 思考过程:")
+        lines.append(f"```\n{result.thinking}\n```\n")
+
+        if result.action:
+            lines.append(f"### 🎯 执行动作:")
+            lines.append(f"```json\n{json.dumps(result.action, ensure_ascii=False, indent=2)}\n```\n")
+
+        if result.message:
+            lines.append(f"### 📝 消息:")
+            lines.append(f"> {result.message}\n")
+
+        return "\n".join(lines)
+
+    def _format_step_output(self, title: str, content: str, level: str = "info") -> str:
+        """Format a step output message."""
+        icons = {
+            "info": "ℹ️",
+            "success": "✅",
+            "warning": "⚠️",
+            "error": "❌",
+        }
+        icon = icons.get(level, "ℹ️")
+
+        lines = [f"\n{'='*60}", f"{icon} **{title}**", f"{'='*60}\n"]
+        lines.append(content)
+        return "\n".join(lines)
+
+    def _get_screenshot(self) -> str:
+        """Get current device screenshot and save to temp file."""
+        try:
+            import base64
+            from io import BytesIO
+            from PIL import Image
+
+            screenshot = get_screenshot(self.agent_config.device_id)
+            # Convert base64 data back to image
+            img_data = base64.b64decode(screenshot.base64_data)
+            img = Image.open(BytesIO(img_data))
+
+            # Save to temporary file
+            temp_path = os.path.join(tempfile.gettempdir(), f"phone_agent_screenshot_{int(time.time())}.png")
+            img.save(temp_path)
+            return temp_path
+        except Exception as e:
+            print(f"Screenshot error: {e}")
+            return None
+
+
+
+# Global session storage
+sessions = {}
+current_session_id = 0
+
+
+def get_or_create_session(base_url: str, model_name: str, api_key: str,
+                          max_steps: int, device_id: str, lang: str,
+                          session_id: int = None) -> Tuple[AgentSession, int]:
+    """Get existing session or create new one."""
+    global current_session_id
+
+    if session_id is None or session_id not in sessions:
+        current_session_id += 1
+        session_id = current_session_id
+        sessions[session_id] = AgentSession(
+            base_url, model_name, api_key, max_steps, device_id, lang
+        )
+
+    return sessions[session_id], session_id
+
+
+# ============================================================================
+# Gradio Interface Functions
+# ============================================================================
+
+def refresh_devices():
+    """Refresh and return device list."""
+    try:
+        devices = list_devices()
+        if not devices:
+            return gr.Dropdown(choices=[], value=None, label="可用设备 (无设备)")
+
+        device_choices = [f"{d.device_id} ({d.model or 'Unknown'})" for d in devices]
+        device_ids = [d.device_id for d in devices]
+
+        return gr.Dropdown(
+            choices=list(zip(device_choices, device_ids)),
+            value=device_ids[0] if device_ids else None,
+            label=f"可用设备 ({len(devices)} 个)"
+        )
+    except Exception as e:
+        return gr.Dropdown(choices=[], value=None, label=f"设备检查失败: {str(e)}")
+
+
+def run_system_check(base_url, model_name, api_key, device_id):
+    """Run system check and return report."""
+    return run_full_system_check(base_url, model_name, api_key, device_id)
+
+
+def execute_task(task, base_url, model_name, api_key, max_steps, device_id, lang, session_id):
+    """Execute a task with streaming output."""
+    if not task or not task.strip():
+        yield "请输入任务内容", None, session_id
+        return
+
+    session, new_session_id = get_or_create_session(
+        base_url, model_name, api_key, max_steps, device_id, lang, session_id
+    )
+
+    for output, screenshot in session.run_task_stream(task.strip()):
+        yield output, screenshot, new_session_id
+
+
+def new_conversation(base_url, model_name, api_key, max_steps, device_id, lang):
+    """Start a new conversation."""
+    global current_session_id
+    current_session_id += 1
+    session_id = current_session_id
+
+    sessions[session_id] = AgentSession(
+        base_url, model_name, api_key, max_steps, device_id, lang
+    )
+
+    return "开始新对话", None, session_id
+
+
+def get_current_screenshot(device_id):
+    """Get and display current device screenshot."""
+    try:
+        import base64
+        from io import BytesIO
+        from PIL import Image
+
+        screenshot = get_screenshot(device_id)
+        # Convert base64 data back to image
+        img_data = base64.b64decode(screenshot.base64_data)
+        img = Image.open(BytesIO(img_data))
+
+        # Save to temp path
+        temp_path = os.path.join(tempfile.gettempdir(), f"phone_agent_screenshot_current_{int(time.time())}.png")
+        img.save(temp_path)
+        return temp_path
+    except Exception as e:
+        print(f"Screenshot error: {e}")
+        return None
+
+
+def stop_current_task(session_id):
+    """Stop the current running task."""
+    if session_id and session_id in sessions:
+        sessions[session_id].stop_task()
+        return "正在终止任务..."
+    return "没有运行中的任务"
+
+
+
+
+# ============================================================================
+# Internationalization (i18n)
+# ============================================================================
+
+I18N = {
+    "en": {
+        "title": "# 🤖 Phone Agent\n\nAI-powered phone automation assistant - Control your Android device with natural language",
+        "device_management": "## 📱 Device Management",
+        "device_management_desc": "View and manage connected Android devices",
+        "wifi_connection": "**WiFi Wireless Connection**",
+        "device_ip_label": "Device IP Address",
+        "device_ip_placeholder": "e.g., 192.168.1.100:5555",
+        "wifi_connect_btn": "📡 WiFi Connect",
+        "select_device": "Select Device",
+        "refresh_devices": "🔄 Refresh Devices",
+        "list_devices": "📋 List Devices",
+        "device_output_init": "Click the button above to view device information or connect devices",
+        "system_check": "## 🔍 System Check",
+        "system_check_desc": "Click the button below to check if the system environment meets the requirements",
+        "run_check_btn": "▶️ Run System Check",
+        "check_output_init": "Click the button above to start checking",
+        "conversation_control": "## 💬 Conversation Control",
+        "model_api_url": "Model API URL",
+        "model_name": "Model Name",
+        "api_key": "API Key",
+        "max_steps": "Max Steps",
+        "task_input": "📝 Enter Task",
+        "task_placeholder": "e.g., Open WeChat and send a message 'Hello' to John",
+        "execute_task": "🚀 Execute Task",
+        "stop_task": "⏹️ Stop Task",
+        "new_conversation": "➕ New Conversation",
+        "execution_output": "📤 Execution Output",
+        "output_init": "Waiting for task input...",
+        "device_screenshot": "📱 Device Screenshot",
+        "auto_refresh": "Auto Refresh (0.5s)",
+        "help": "## ❓ Help",
+        "user_guide": "📖 User Guide",
+    },
+    "cn": {
+        "title": "# 🤖 Phone Agent 手机助手\n\n基于AI的智能手机自动化助手 - 通过自然语言控制您的Android设备",
+        "device_management": "## 📱 设备管理",
+        "device_management_desc": "查看和管理已连接的Android设备",
+        "wifi_connection": "**WiFi无线连接**",
+        "device_ip_label": "设备IP地址",
+        "device_ip_placeholder": "例如: 192.168.1.100:5555",
+        "wifi_connect_btn": "📡 WiFi连接",
+        "select_device": "选择设备",
+        "refresh_devices": "🔄 刷新设备",
+        "list_devices": "📋 列出设备",
+        "device_output_init": "点击上方按钮查看设备信息或连接设备",
+        "system_check": "## 🔍 系统检查",
+        "system_check_desc": "点击下方按钮检查系统环境是否满足运行要求",
+        "run_check_btn": "▶️ 运行系统检查",
+        "check_output_init": "点击上方按钮开始检查",
+        "conversation_control": "## 💬 对话控制",
+        "model_api_url": "模型API地址",
+        "model_name": "模型名称",
+        "api_key": "API Key",
+        "max_steps": "最大步数",
+        "task_input": "📝 输入任务",
+        "task_placeholder": "例如: 打开微信,给张三发送一条消息'你好'",
+        "execute_task": "🚀 执行任务",
+        "stop_task": "⏹️ 终止任务",
+        "new_conversation": "➕ 新对话",
+        "execution_output": "📤 执行输出",
+        "output_init": "等待任务输入...",
+        "device_screenshot": "📱 设备截图",
+        "auto_refresh": "自动刷新(0.5s)",
+        "help": "## ❓ 帮助",
+        "user_guide": "📖 使用指南",
+    }
+}
+
+# ============================================================================
+# Gradio UI
+# ============================================================================
+
+def create_ui():
+    """Create the Gradio interface."""
+
+    # Default language from environment variable
+    default_lang = os.getenv("PHONE_AGENT_LANG", "en")
+
+    with gr.Blocks(title="Phone Agent", theme=gr.themes.Soft()) as app:
+        # Language selector at the top right
+        with gr.Row():
+            with gr.Column(scale=4):
+                title_md = gr.Markdown(I18N[default_lang]["title"])
+            with gr.Column(scale=1):
+                lang = gr.Radio(
+                    choices=[("English", "en"), ("简体中文", "cn")],
+                    value=default_lang,
+                    label="Language / 语言",
+                    info="Agent language"
+                )
+
+        # Session state
+        session_state = gr.State(None)
+
+        # ================================================================
+        # Section 1: Device Management
+        # ================================================================
+        device_mgmt_title = gr.Markdown(I18N[default_lang]["device_management"])
+        device_mgmt_desc = gr.Markdown(I18N[default_lang]["device_management_desc"])
+
+        wifi_connection_md = gr.Markdown(I18N[default_lang]["wifi_connection"])
+
+        with gr.Row():
+            wifi_ip = gr.Textbox(
+                label=I18N[default_lang]["device_ip_label"],
+                placeholder=I18N[default_lang]["device_ip_placeholder"],
+            )
+            wifi_connect_btn = gr.Button(I18N[default_lang]["wifi_connect_btn"], size="lg", variant="primary")
+
+        with gr.Row():
+            device_dropdown = gr.Dropdown(
+                label=I18N[default_lang]["select_device"],
+                choices=[],
+                value=None,
+                interactive=True,
+                scale=2
+            )
+            refresh_btn = gr.Button(I18N[default_lang]["refresh_devices"], size="lg")
+            list_devices_btn = gr.Button(I18N[default_lang]["list_devices"], size="lg")
+
+        device_output = gr.Markdown(I18N[default_lang]["device_output_init"])
+
+        gr.Markdown("---")  # Divider
+
+        # ================================================================
+        # Section 2: System Check
+        # ================================================================
+        system_check_title = gr.Markdown(I18N[default_lang]["system_check"])
+        system_check_desc = gr.Markdown(I18N[default_lang]["system_check_desc"])
+
+        check_btn = gr.Button(I18N[default_lang]["run_check_btn"], variant="primary", size="lg")
+        check_output = gr.Markdown(I18N[default_lang]["check_output_init"])
+
+        gr.Markdown("---")  # Divider
+
+        # ================================================================
+        # Section 3: Main Interface - Conversation Control
+        # ================================================================
+        conversation_title = gr.Markdown(I18N[default_lang]["conversation_control"])
+
+        with gr.Row():
+            with gr.Column(scale=2):
+                # Configuration Section
+                with gr.Row():
+                    base_url = gr.Textbox(
+                        label=I18N[default_lang]["model_api_url"],
+                        value=os.getenv("PHONE_AGENT_BASE_URL", "http://localhost:8000/v1"),
+                        placeholder="http://localhost:8000/v1"
+                    )
+                    model_name = gr.Textbox(
+                        label=I18N[default_lang]["model_name"],
+                        value=os.getenv("PHONE_AGENT_MODEL", "autoglm-phone-9b"),
+                        placeholder="autoglm-phone-9b"
+                    )
+
+                with gr.Row():
+                    api_key = gr.Textbox(
+                        label=I18N[default_lang]["api_key"],
+                        value=os.getenv("PHONE_AGENT_API_KEY", "EMPTY"),
+                        placeholder="EMPTY",
+                        type="password"
+                    )
+                    max_steps = gr.Number(
+                        label=I18N[default_lang]["max_steps"],
+                        value=int(os.getenv("PHONE_AGENT_MAX_STEPS", "100")),
+                        precision=0
+                    )
+
+                # Task Input
+                task_input = gr.Textbox(
+                    label=I18N[default_lang]["task_input"],
+                    placeholder=I18N[default_lang]["task_placeholder"],
+                    lines=3
+                )
+
+                with gr.Row():
+                    submit_btn = gr.Button(I18N[default_lang]["execute_task"], variant="primary", size="lg")
+                    stop_btn = gr.Button(I18N[default_lang]["stop_task"], variant="stop", size="lg")
+                    new_chat_btn = gr.Button(I18N[default_lang]["new_conversation"], size="lg")
+
+                # Output Display
+                output_display = gr.Markdown(
+                    label=I18N[default_lang]["execution_output"],
+                    value=I18N[default_lang]["output_init"]
+                )
+
+            with gr.Column(scale=1):
+                # Screenshot Display
+                screenshot_display = gr.Image(
+                    label=I18N[default_lang]["device_screenshot"],
+                    type="filepath",
+                    height=800,
+                    width=360
+
+                )
+
+                auto_refresh_checkbox = gr.Checkbox(
+                    label=I18N[default_lang]["auto_refresh"],
+                    value=False
+                )
+
+                # Auto-refresh timer
+                screenshot_timer = gr.Timer(value=0.5, active=False)
+
+        gr.Markdown("---")  # Divider
+
+        # ================================================================
+        # Section 4: Help
+        # ================================================================
+        help_title = gr.Markdown(I18N[default_lang]["help"])
+
+        with gr.Accordion(I18N[default_lang]["user_guide"], open=False):
+            gr.Markdown("""
+            ### 快速开始
+
+            1. **检查环境**: 在"系统检查"部分，运行系统检查确保环境配置正确
+            2. **配置设置**: 在"配置"区域设置模型API地址和API Key
+            3. **选择设备**: 点击"刷新设备"按钮，从下拉列表中选择目标设备
+            4. **输入任务**: 在任务输入框中用自然语言描述您的需求
+            5. **执行任务**: 点击"执行任务"按钮，系统会自动控制手机完成任务
+
+            ### 功能说明
+
+            #### 💬 对话控制
+            - **任务执行**: 支持流式输出，实时显示AI的思考过程和执行动作
+            - **截图显示**: 右侧实时显示设备当前屏幕状态
+            - **新对话**: 清空当前上下文，开始全新的任务会话
+
+            #### 🔍 系统检查
+            - 检查ADB工具安装状态
+            - 检查设备连接状态
+            - 检查ADB Keyboard安装状态
+            - 检查模型API连接状态
+
+            #### 📱 设备管理
+            - 查看所有已连接设备
+            - 支持USB和WiFi连接
+            - 远程设备连接功能
+
+            ### 任务示例
+
+            ```
+            # 消息发送
+            打开微信，给张三发送消息"晚上一起吃饭吗?"
+
+            # 应用操作
+            打开抖音，搜索"美食教程"，点赞第一个视频
+
+            # 购物任务
+            打开淘宝，搜索"机械键盘"，按价格从低到高排序，加购第一个商品
+
+            # 信息查询
+            打开小红书，搜索"成都旅游攻略"，总结前5篇笔记的内容
+            ```
+
+            ### 注意事项
+
+            - 首次使用请确保已安装ADB Keyboard并在设备设置中启用
+            - 执行任务前请确保设备已解锁
+            - 某些敏感操作(如支付)可能需要人工确认
+            - 建议在WiFi环境下使用以获得更好的响应速度
+
+            ### 常见问题
+
+            **Q: 无法检测到设备?**
+            A: 确保已启用USB调试，并在设备上授权计算机的调试请求
+
+            **Q: 任务执行失败?**
+            A: 检查网络连接，确认模型API服务正常运行
+
+            **Q: 输入文本没有反应?**
+            A: 确保已安装并启用ADB Keyboard
+
+            ### 技术支持
+
+            - GitHub: [Open-AutoGLM](https://github.com/zai-org/Open-AutoGLM)
+            - 文档: 查看项目README获取更多信息
+            """)
+
+        # ================================================================
+        # Event Handlers
+        # ================================================================
+
+        # Language change handler
+        def update_language(selected_lang):
+            """Update all UI text based on selected language."""
+            i18n = I18N.get(selected_lang, I18N["en"])
+            return [
+                # Title
+                i18n["title"],
+                # Device Management
+                i18n["device_management"],
+                i18n["device_management_desc"],
+                i18n["wifi_connection"],
+                gr.Textbox(label=i18n["device_ip_label"], placeholder=i18n["device_ip_placeholder"]),
+                gr.Button(value=i18n["wifi_connect_btn"]),
+                gr.Dropdown(label=i18n["select_device"]),
+                gr.Button(value=i18n["refresh_devices"]),
+                gr.Button(value=i18n["list_devices"]),
+                # System Check
+                i18n["system_check"],
+                i18n["system_check_desc"],
+                gr.Button(value=i18n["run_check_btn"]),
+                # Conversation Control
+                i18n["conversation_control"],
+                gr.Textbox(label=i18n["model_api_url"]),
+                gr.Textbox(label=i18n["model_name"]),
+                gr.Textbox(label=i18n["api_key"]),
+                gr.Number(label=i18n["max_steps"]),
+                gr.Textbox(label=i18n["task_input"], placeholder=i18n["task_placeholder"]),
+                gr.Button(value=i18n["execute_task"]),
+                gr.Button(value=i18n["stop_task"]),
+                gr.Button(value=i18n["new_conversation"]),
+                gr.Image(label=i18n["device_screenshot"]),
+                gr.Checkbox(label=i18n["auto_refresh"]),
+                # Help
+                i18n["help"],
+            ]
+
+        # Language selector change event
+        lang.change(
+            fn=update_language,
+            inputs=[lang],
+            outputs=[
+                title_md,
+                device_mgmt_title, device_mgmt_desc,
+                wifi_connection_md,
+                wifi_ip, wifi_connect_btn,
+                device_dropdown, refresh_btn, list_devices_btn,
+                system_check_title, system_check_desc, check_btn,
+                conversation_title,
+                base_url, model_name, api_key, max_steps,
+                task_input, submit_btn, stop_btn, new_chat_btn,
+                screenshot_display, auto_refresh_checkbox,
+                help_title,
+            ]
+        )
+
+        # Refresh devices
+        refresh_btn.click(
+            fn=refresh_devices,
+            inputs=[],
+            outputs=[device_dropdown]
+        )
+
+        # System check
+        check_btn.click(
+            fn=run_system_check,
+            inputs=[base_url, model_name, api_key, device_dropdown],
+            outputs=[check_output]
+        )
+
+        # Execute task
+        submit_btn.click(
+            fn=execute_task,
+            inputs=[task_input, base_url, model_name, api_key, max_steps,
+                   device_dropdown, lang, session_state],
+            outputs=[output_display, screenshot_display, session_state]
+        )
+
+        # New conversation
+        new_chat_btn.click(
+            fn=new_conversation,
+            inputs=[base_url, model_name, api_key, max_steps, device_dropdown, lang],
+            outputs=[output_display, screenshot_display, session_state]
+        )
+
+        # Stop task
+        stop_btn.click(
+            fn=stop_current_task,
+            inputs=[session_state],
+            outputs=[output_display]
+        )
+
+        # List devices
+        def list_devices_info():
+            try:
+                devices = list_devices()
+                if not devices:
+                    return "## 设备列表\n\n❌ 没有检测到连接的设备"
+
+                lines = ["## 设备列表\n"]
+                for i, device in enumerate(devices, 1):
+                    lines.append(f"### {i}. {device.device_id}")
+                    lines.append(f"- **型号**: {device.model or 'Unknown'}")
+                    lines.append(f"- **状态**: {device.status}")
+                    lines.append(f"- **连接类型**: {device.connection_type.value}")
+                    lines.append("")
+
+                return "\n".join(lines)
+            except Exception as e:
+                return f"## 设备列表\n\n❌ 错误: {str(e)}"
+
+        list_devices_btn.click(
+            fn=list_devices_info,
+            inputs=[],
+            outputs=[device_output]
+        )
+
+        # WiFi Connect device
+        def wifi_connect_device(ip_address):
+            if not ip_address:
+                return "❌ 请输入设备IP地址"
+
+            try:
+                from phone_agent.adb import ADBConnection
+                conn = ADBConnection()
+                success, message = conn.connect(ip_address)
+                if success:
+                    return f"✅ WiFi连接成功: {message}\n\n提示: 连接成功后请点击'刷新设备'按钮更新设备列表"
+                else:
+                    return f"❌ WiFi连接失败: {message}"
+            except Exception as e:
+                return f"❌ WiFi连接失败: {str(e)}"
+
+        wifi_connect_btn.click(
+            fn=wifi_connect_device,
+            inputs=[wifi_ip],
+            outputs=[device_output]
+        )
+
+        # Auto-refresh screenshot
+        # Toggle timer active state based on checkbox
+        auto_refresh_checkbox.change(
+            fn=lambda checked: gr.Timer(active=checked),
+            inputs=[auto_refresh_checkbox],
+            outputs=[screenshot_timer]
+        )
+
+        # Update screenshot on timer tick
+        screenshot_timer.tick(
+            fn=get_current_screenshot,
+            inputs=[device_dropdown],
+            outputs=[screenshot_display]
+        )
+
+    return app
+
+
+# ============================================================================
+# Main Entry Point
+# ============================================================================
+
+def main():
+    """Launch the Gradio app."""
+    app = create_ui()
+    app.launch(
+        server_name="127.0.0.1",
+        server_port=7860,
+        share=False,
+        show_error=True,
+        inbrowser=False,
+        quiet=False,
+    )
+
+
+if __name__ == "__main__":
+    main()