chore: improve documentation for extra_user_content_parts in Provider classes

fixes: #4176

.github/workflows/build-app.yml

@@ -1,79 +0,0 @@
-name: Build Desktop App
-
-on:
-  push:
-    tags:
-      - 'v*'
-  workflow_dispatch:
-
-jobs:
-  build:
-    strategy:
-      fail-fast: false
-      matrix:
-        platform: [macos-latest, ubuntu-latest, windows-latest]
-
-    runs-on: ${{ matrix.platform }}
-    
-    steps:
-      - uses: actions/checkout@v4
-
-      - name: Setup Python
-        uses: actions/setup-python@v5
-        with:
-          python-version: '3.10'
-
-      - name: Setup Node.js
-        uses: actions/setup-node@v4
-        with:
-          node-version: 20
-
-      - name: Install Rust
-        uses: dtolnay/rust-toolchain@stable
-
-      - name: Install dependencies (Ubuntu)
-        if: matrix.platform == 'ubuntu-latest'
-        run: |
-          sudo apt-get update
-          sudo apt-get install -y libgtk-3-dev libwebkit2gtk-4.0-dev libappindicator3-dev librsvg2-dev patchelf
-
-      - name: Install Python dependencies
-        run: |
-          pip install uv
-          uv sync
-
-      - name: Build Python backend with Nuitka
-        run: |
-          pip install nuitka
-          python build_nuitka.py
-
-      - name: Install Node dependencies
-        working-directory: ./dashboard
-        run: npm install
-
-      - name: Build Tauri app
-        working-directory: ./dashboard
-        run: npm run tauri:build
-
-      - name: Upload artifacts (macOS)
-        if: matrix.platform == 'macos-latest'
-        uses: actions/upload-artifact@v4
-        with:
-          name: astrbot-macos
-          path: dashboard/src-tauri/target/release/bundle/dmg/*.dmg
-
-      - name: Upload artifacts (Windows)
-        if: matrix.platform == 'windows-latest'
-        uses: actions/upload-artifact@v4
-        with:
-          name: astrbot-windows
-          path: dashboard/src-tauri/target/release/bundle/msi/*.msi
-
-      - name: Upload artifacts (Linux)
-        if: matrix.platform == 'ubuntu-latest'
-        uses: actions/upload-artifact@v4
-        with:
-          name: astrbot-linux
-          path: |
-            dashboard/src-tauri/target/release/bundle/deb/*.deb
-            dashboard/src-tauri/target/release/bundle/appimage/*.AppImage

.gitignore

@@ -32,7 +32,6 @@ tests/astrbot_plugin_openai
 # Dashboard
 dashboard/node_modules/
 dashboard/dist/
-dashboard/src-tauri/target
 package-lock.json
 package.json
 yarn.lock
@@ -49,6 +48,5 @@ astrbot.lock
 chroma
 venv/*
 pytest.ini
-build/
 AGENTS.md
 IFLOW.md

BUILD_INSTRUCTIONS.md

@@ -1,287 +0,0 @@
-# AstrBot 桌面应用构建指南
-
-本指南介绍如何使用 Nuitka 将 Python 后端打包并集成到 Tauri 桌面应用中。
-
-## 前置要求
-
-### 系统要求
-- Python 3.10+
-- Node.js 20+
-- Rust (通过 rustup 安装)
-- UV 包管理器
-
-### macOS 额外要求
-- Xcode Command Line Tools: `xcode-select --install`
-
-### Linux 额外要求
-```bash
-sudo apt-get install -y libgtk-3-dev libwebkit2gtk-4.0-dev \
-  libappindicator3-dev librsvg2-dev patchelf
-```
-
-### Windows 额外要求
-- Visual Studio 2019+ with C++ build tools
-- Windows 10 SDK
-
-## 构建步骤
-
-### 1. 安装 Python 依赖
-```bash
-pip install uv
-uv sync
-```
-
-### 2. 安装 Nuitka
-```bash
-pip install nuitka
-```
-
-### 3. 构建 Python 后端
-```bash
-python build_nuitka.py
-```
-
-这会使用 Nuitka 将 `main.py` 编译为独立可执行文件，输出到 `build/nuitka/` 目录。
-
-**注意**: Nuitka 编译过程可能需要 10-30 分钟，取决于您的系统性能。
-
-### 4. 安装前端依赖
-```bash
-cd dashboard
-npm install
-```
-
-### 5. 构建 Tauri 应用
-```bash
-npm run tauri:build
-```
-
-构建脚本会自动:
-1. 运行 `build_nuitka.py` 编译 Python 后端
-2. 将编译好的可执行文件复制到 `src-tauri/resources/` 目录
-3. 构建 Tauri 应用并打包所有资源
-
-### 6. 查找构建产物
-
-构建完成后，您可以在以下位置找到安装包:
-
-- **macOS**: `dashboard/src-tauri/target/release/bundle/dmg/AstrBot_*.dmg`
-- **Windows**: `dashboard/src-tauri/target/release/bundle/msi/AstrBot_*.msi`
-- **Linux**: 
-  - `dashboard/src-tauri/target/release/bundle/deb/astrbot_*.deb`
-  - `dashboard/src-tauri/target/release/bundle/appimage/astrbot_*.AppImage`
-
-## 开发模式
-
-在开发时，您可能不想每次都完整编译 Python 后端。
-
-### 仅开发 Tauri + Vue
-```bash
-cd dashboard
-npm run tauri:dev
-```
-
-这会启动开发服务器，但不会自动启动 Python 后端。您需要手动运行:
-```bash
-uv run main.py
-```
-
-### 测试完整集成
-如果您想测试 Tauri 自动启动 Python 后端的功能:
-
-1. 先编译一次 Python 后端:
-```bash
-python build_nuitka.py
-```
-
-2. 手动复制到资源目录:
-```bash
-# macOS
-cp -r build/nuitka/main.app dashboard/src-tauri/resources/astrbot-backend.app
-
-# Windows
-copy build\nuitka\main.exe dashboard\src-tauri\resources\astrbot-backend.exe
-
-# Linux
-cp build/nuitka/main.bin dashboard/src-tauri/resources/astrbot-backend
-```
-
-3. 运行开发模式:
-```bash
-cd dashboard
-npm run tauri:dev
-```
-
-## Nuitka 构建选项说明
-
-`build_nuitka.py` 脚本使用以下关键选项:
-
-- `--standalone`: 创建包含所有依赖的独立目录
-- `--onefile`: 将所有内容打包到单个可执行文件
-- `--follow-imports`: 自动跟踪所有 Python 导入
-- `--include-package`: 明确包含特定包
-- `--include-data-dir`: 包含数据目录（插件、配置等）
-
-### 自定义构建
-
-如果您需要修改构建选项，编辑 `build_nuitka.py`:
-
-```python
-# 添加更多要包含的包
-include_packages = [
-    "astrbot",
-    "your_custom_package",
-    # ...
-]
-
-# 添加更多数据目录
-data_includes = [
-    "data/config",
-    "your_custom_data",
-    # ...
-]
-```
-
-## 常见问题
-
-### 1. Nuitka 编译失败
-**问题**: 编译时出现 "module not found" 错误
-
-**解决方案**: 在 `build_nuitka.py` 中添加缺失的包到 `include_packages` 列表
-
-### 2. 运行时找不到资源文件
-**问题**: 应用启动后提示找不到配置文件或插件
-
-**解决方案**: 确保在 `build_nuitka.py` 中使用 `--include-data-dir` 包含了所有必要的数据目录
-
-### 3. macOS 安全警告
-**问题**: macOS 提示"应用来自未知开发者"
-
-**解决方案**: 
-```bash
-# 临时解除限制
-sudo spctl --master-disable
-
-# 或者为特定应用授权
-xattr -cr /Applications/AstrBot.app
-```
-
-对于生产发布，您需要:
-1. 注册 Apple Developer 账号
-2. 对应用进行代码签名
-3. 提交公证 (Notarization)
-
-### 4. Windows Defender 报毒
-**问题**: Windows Defender 或其他杀毒软件报毒
-
-**解决方案**: 
-- 这是 Nuitka 打包程序的常见问题
-- 可以使用 `--windows-company-name` 和 `--windows-product-name` 添加元数据
-- 对于生产发布，需要购买代码签名证书
-
-### 5. Linux 依赖问题
-**问题**: 在某些 Linux 发行版上缺少共享库
-
-**解决方案**: 使用 AppImage 格式，它包含所有依赖:
-```bash
-# 构建时会自动生成 AppImage
-npm run tauri:build
-```
-
-## 优化构建大小
-
-默认的 `--onefile` 模式会生成较大的可执行文件。如果需要减小体积:
-
-1. 移除不需要的包
-2. 使用 `--standalone` 而不是 `--onefile`
-3. 排除不必要的数据文件
-
-修改 `build_nuitka.py`:
-```python
-# 移除 --onefile，使用 --standalone
-nuitka_cmd = [
-    sys.executable,
-    "-m", "nuitka",
-    "--standalone",  # 只使用 standalone
-    # "--onefile",   # 注释掉 onefile
-    # ...
-]
-```
-
-## CI/CD 集成
-
-项目已配置 GitHub Actions 工作流 (`.github/workflows/build-app.yml`)，可以自动为所有平台构建应用。
-
-推送标签时自动触发:
-```bash
-git tag v4.5.7
-git push origin v4.5.7
-```
-
-或手动触发:
-在 GitHub Actions 页面选择 "Build Desktop App" 工作流并点击 "Run workflow"
-
-## 发布清单
-
-在发布新版本前:
-
-- [ ] 更新版本号
-  - `pyproject.toml` - Python 项目版本
-  - `dashboard/package.json` - Node 项目版本
-  - `dashboard/src-tauri/Cargo.toml` - Rust 项目版本
-  - `dashboard/src-tauri/tauri.conf.json` - Tauri 配置版本
-
-- [ ] 运行代码检查
-  ```bash
-  uv run ruff check .
-  uv run ruff format .
-  ```
-
-- [ ] 本地测试构建
-  ```bash
-  python build_nuitka.py
-  cd dashboard && npm run tauri:build
-  ```
-
-- [ ] 测试安装包
-  - 安装生成的安装包
-  - 验证应用启动
-  - 验证 Python 后端自动启动
-  - 测试核心功能
-
-- [ ] 创建发布标签
-  ```bash
-  git tag -a v4.5.7 -m "Release v4.5.7"
-  git push origin v4.5.7
-  ```
-
-## 技术架构
-
-```
-┌─────────────────────────────────────┐
-│         Tauri Desktop App           │
-│  (Rust + WebView)                   │
-│                                     │
-│  ┌─────────────────────────────┐   │
-│  │   Vue.js Dashboard          │   │
-│  │   (Frontend UI)             │   │
-│  └─────────────────────────────┘   │
-│                                     │
-│  ┌─────────────────────────────┐   │
-│  │   Python Backend            │   │
-│  │   (Nuitka Compiled)         │   │
-│  │   - AstrBot Core            │   │
-│  │   - Plugins                 │   │
-│  │   - API Server              │   │
-│  └─────────────────────────────┘   │
-│                                     │
-│         HTTP/WebSocket              │
-│      localhost:6185                 │
-└─────────────────────────────────────┘
-```
-
-## 参考资源
-
-- [Nuitka 文档](https://nuitka.net/doc/user-manual.html)
-- [Tauri 文档](https://tauri.app/v1/guides/)
-- [AstrBot 文档](https://astrbot.fun)

README.md

@@ -1,4 +1,4 @@
-![AstrBot-Logo-Simplified](https://github.com/user-attachments/assets/ffd99b6b-3272-4682-beaa-6fe74250f7d9)
+![astrbot-banner-xmas](https://github.com/user-attachments/assets/bf2341de-ec7a-45a7-a04a-02ad36450e99)
 
 <div align="center">
 

astrbot/cli/__init__.py

@@ -1 +1 @@
-__version__ = "4.9.2"
+__version__ = "4.10.2"

astrbot/core/agent/runners/tool_loop_agent_runner.py

@@ -76,12 +76,20 @@ class ToolLoopAgentRunner(BaseAgentRunner[TContext]):
 
     async def _iter_llm_responses(self) -> T.AsyncGenerator[LLMResponse, None]:
         """Yields chunks *and* a final LLMResponse."""
+        payload = {
+            "contexts": self.run_context.messages,  # list[Message]
+            "func_tool": self.req.func_tool,
+            "model": self.req.model,  # NOTE: in fact, this arg is None in most cases
+            "session_id": self.req.session_id,
+            "extra_user_content_parts": self.req.extra_user_content_parts,  # list[ContentPart]
+        }
+
         if self.streaming:
-            stream = self.provider.text_chat_stream(**self.req.__dict__)
+            stream = self.provider.text_chat_stream(**payload)
             async for resp in stream:  # type: ignore
                 yield resp
         else:
-            yield await self.provider.text_chat(**self.req.__dict__)
+            yield await self.provider.text_chat(**payload)
 
     @override
     async def step(self):
@@ -165,7 +173,7 @@ class ToolLoopAgentRunner(BaseAgentRunner[TContext]):
             self.run_context.messages.append(
                 Message(
                     role="assistant",
-                    content=llm_resp.completion_text or "",
+                    content=llm_resp.completion_text or "*No response*",
                 ),
             )
             try:
@@ -230,6 +238,25 @@ class ToolLoopAgentRunner(BaseAgentRunner[TContext]):
             async for resp in self.step():
                 yield resp
 
+        #  如果循环结束了但是 agent 还没有完成，说明是达到了 max_step
+        if not self.done():
+            logger.warning(
+                f"Agent reached max steps ({max_step}), forcing a final response."
+            )
+            # 拔掉所有工具
+            if self.req:
+                self.req.func_tool = None
+            # 注入提示词
+            self.run_context.messages.append(
+                Message(
+                    role="user",
+                    content="工具调用次数已达到上限，请停止使用工具，并根据已经收集到的信息，对你的任务和发现进行总结，然后直接回复用户。",
+                )
+            )
+            # 再执行最后一步
+            async for resp in self.step():
+                yield resp
+
     async def _handle_function_tools(
         self,
         req: ProviderRequest,
@@ -376,35 +403,33 @@ class ToolLoopAgentRunner(BaseAgentRunner[TContext]):
                                     ),
                                 )
 
-                        # yield the last tool call result
-                        if tool_call_result_blocks:
-                            last_tcr_content = str(tool_call_result_blocks[-1].content)
-                            yield MessageChain(
-                                type="tool_call_result",
-                                chain=[
-                                    Json(
-                                        data={
-                                            "id": func_tool_id,
-                                            "ts": time.time(),
-                                            "result": last_tcr_content,
-                                        }
-                                    )
-                                ],
-                            )
-
                     elif resp is None:
                         # Tool 直接请求发送消息给用户
                         # 这里我们将直接结束 Agent Loop。
                         # 发送消息逻辑在 ToolExecutor 中处理了。
                         logger.warning(
-                            f"{func_tool_name} 没有没有返回值或者将结果直接发送给用户，此工具调用不会被记录到历史中。"
+                            f"{func_tool_name} 没有没有返回值或者将结果直接发送给用户。"
                         )
                         self._transition_state(AgentState.DONE)
                         self.stats.end_time = time.time()
+                        tool_call_result_blocks.append(
+                            ToolCallMessageSegment(
+                                role="tool",
+                                tool_call_id=func_tool_id,
+                                content="*工具没有返回值或者将结果直接发送给了用户*",
+                            ),
+                        )
                     else:
                         # 不应该出现其他类型
                         logger.warning(
-                            f"Tool 返回了不支持的类型: {type(resp)}，将忽略。",
+                            f"Tool 返回了不支持的类型: {type(resp)}。",
+                        )
+                        tool_call_result_blocks.append(
+                            ToolCallMessageSegment(
+                                role="tool",
+                                tool_call_id=func_tool_id,
+                                content="*工具返回了不支持的类型，请告诉用户检查这个工具的定义和实现。*",
+                            ),
                         )
 
                 try:
@@ -426,6 +451,22 @@ class ToolLoopAgentRunner(BaseAgentRunner[TContext]):
                     ),
                 )
 
+        # yield the last tool call result
+        if tool_call_result_blocks:
+            last_tcr_content = str(tool_call_result_blocks[-1].content)
+            yield MessageChain(
+                type="tool_call_result",
+                chain=[
+                    Json(
+                        data={
+                            "id": func_tool_id,
+                            "ts": time.time(),
+                            "result": last_tcr_content,
+                        }
+                    )
+                ],
+            )
+
         # 处理函数调用响应
         if tool_call_result_blocks:
             yield tool_call_result_blocks

astrbot/core/astr_agent_run_util.py

@@ -2,6 +2,7 @@ import traceback
 from collections.abc import AsyncGenerator
 
 from astrbot.core import logger
+from astrbot.core.agent.message import Message
 from astrbot.core.agent.runners.tool_loop_agent_runner import ToolLoopAgentRunner
 from astrbot.core.astr_agent_context import AstrAgentContext
 from astrbot.core.message.components import Json
@@ -24,8 +25,25 @@ async def run_agent(
 ) -> AsyncGenerator[MessageChain | None, None]:
     step_idx = 0
     astr_event = agent_runner.run_context.context.event
-    while step_idx < max_step:
+    while step_idx < max_step + 1:
         step_idx += 1
+
+        if step_idx == max_step + 1:
+            logger.warning(
+                f"Agent reached max steps ({max_step}), forcing a final response."
+            )
+            if not agent_runner.done():
+                # 拔掉所有工具
+                if agent_runner.req:
+                    agent_runner.req.func_tool = None
+                # 注入提示词
+                agent_runner.run_context.messages.append(
+                    Message(
+                        role="user",
+                        content="工具调用次数已达到上限，请停止使用工具，并根据已经收集到的信息，对你的任务和发现进行总结，然后直接回复用户。",
+                    )
+                )
+
         try:
             async for resp in agent_runner.step():
                 if astr_event.is_stopped():

astrbot/core/astr_agent_tool_exec.py

@@ -209,12 +209,42 @@ async def call_local_llm_tool(
         else:
             raise ValueError(f"未知的方法名: {method_name}")
     except ValueError as e:
-        logger.error(f"调用本地 LLM 工具时出错: {e}", exc_info=True)
+        raise Exception(f"Tool execution ValueError: {e}") from e
-    except TypeError:
+    except TypeError as e:
-        logger.error("处理函数参数不匹配，请检查 handler 的定义。", exc_info=True)
+        # 获取函数的签名（包括类型），除了第一个 event/context 参数。
+        try:
+            sig = inspect.signature(handler)
+            params = list(sig.parameters.values())
+            # 跳过第一个参数（event 或 context）
+            if params:
+                params = params[1:]
+
+            param_strs = []
+            for param in params:
+                param_str = param.name
+                if param.annotation != inspect.Parameter.empty:
+                    # 获取类型注解的字符串表示
+                    if isinstance(param.annotation, type):
+                        type_str = param.annotation.__name__
+                    else:
+                        type_str = str(param.annotation)
+                    param_str += f": {type_str}"
+                if param.default != inspect.Parameter.empty:
+                    param_str += f" = {param.default!r}"
+                param_strs.append(param_str)
+
+            handler_param_str = (
+                ", ".join(param_strs) if param_strs else "(no additional parameters)"
+            )
+        except Exception:
+            handler_param_str = "(unable to inspect signature)"
+
+        raise Exception(
+            f"Tool handler parameter mismatch, please check the handler definition. Handler parameters: {handler_param_str}"
+        ) from e
     except Exception as e:
         trace_ = traceback.format_exc()
-        logger.error(f"调用本地 LLM 工具时出错: {e}\n{trace_}")
+        raise Exception(f"Tool execution error: {e}. Traceback: {trace_}") from e
 
     if not ready_to_call:
         return

astrbot/core/config/default.py

@@ -1,10 +1,11 @@
 """如需修改配置，请在 `data/cmd_config.json` 中修改或者在管理面板中可视化修改。"""
 
 import os
+from typing import Any, TypedDict
 
 from astrbot.core.utils.astrbot_path import get_astrbot_data_path
 
-VERSION = "4.9.2"
+VERSION = "4.10.2"
 DB_PATH = os.path.join(get_astrbot_data_path(), "data_v4.db")
 
 WEBHOOK_SUPPORTED_PLATFORMS = [
@@ -61,7 +62,8 @@ DEFAULT_CONFIG = {
         "ignore_bot_self_message": False,
         "ignore_at_all": False,
     },
-    "provider": [],
+    "provider_sources": [],  # provider sources
+    "provider": [],  # models from provider_sources
     "provider_settings": {
         "enable": True,
         "default_provider_id": "",
@@ -171,6 +173,22 @@ DEFAULT_CONFIG = {
 }
 
 
+class ChatProviderTemplate(TypedDict):
+    id: str
+    provider_source_id: str
+    model: str
+    modalities: list
+    custom_extra_body: dict[str, Any]
+
+
+CHAT_PROVIDER_TEMPLATE = {
+    "id": "",
+    "provide_source_id": "",
+    "model": "",
+    "modalities": [],
+    "custom_extra_body": {},
+}
+
 """
 AstrBot v3 时代的配置元数据，目前仅承担以下功能：
 
@@ -844,6 +862,7 @@ CONFIG_METADATA_2 = {
         "metadata": {
             "provider": {
                 "type": "list",
+                # provider sources templates
                 "config_template": {
                     "OpenAI": {
                         "id": "openai",
@@ -854,107 +873,10 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.openai.com/v1",
                         "timeout": 120,
-                        "model_config": {"model": "gpt-4o-mini", "temperature": 0.4},
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                        "hint": "也兼容所有与 OpenAI API 兼容的服务。",
                     },
-                    "Azure OpenAI": {
+                    "Google Gemini": {
-                        "id": "azure",
+                        "id": "google_gemini",
-                        "provider": "azure",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "api_version": "2024-05-01-preview",
-                        "key": [],
-                        "api_base": "",
-                        "timeout": 120,
-                        "model_config": {"model": "gpt-4o-mini", "temperature": 0.4},
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "xAI": {
-                        "id": "xai",
-                        "provider": "xai",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": [],
-                        "api_base": "https://api.x.ai/v1",
-                        "timeout": 120,
-                        "model_config": {"model": "grok-2-latest", "temperature": 0.4},
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "xai_native_search": False,
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "Anthropic": {
-                        "hint": "注意Claude系列模型的温度调节范围为0到1.0，超出可能导致报错",
-                        "id": "claude",
-                        "provider": "anthropic",
-                        "type": "anthropic_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": [],
-                        "api_base": "https://api.anthropic.com/v1",
-                        "timeout": 120,
-                        "model_config": {
-                            "model": "claude-3-5-sonnet-latest",
-                            "max_tokens": 4096,
-                            "temperature": 0.2,
-                        },
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "Ollama": {
-                        "hint": "启用前请确保已正确安装并运行 Ollama 服务端，Ollama默认不带鉴权，无需修改key",
-                        "id": "ollama_default",
-                        "provider": "ollama",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": ["ollama"],  # ollama 的 key 默认是 ollama
-                        "api_base": "http://localhost:11434/v1",
-                        "model_config": {"model": "llama3.1-8b", "temperature": 0.4},
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "LM Studio": {
-                        "id": "lm_studio",
-                        "provider": "lm_studio",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": ["lmstudio"],
-                        "api_base": "http://localhost:1234/v1",
-                        "model_config": {
-                            "model": "llama-3.1-8b",
-                        },
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "Gemini(OpenAI兼容)": {
-                        "id": "gemini_default",
-                        "provider": "google",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": [],
-                        "api_base": "https://generativelanguage.googleapis.com/v1beta/openai/",
-                        "timeout": 120,
-                        "model_config": {
-                            "model": "gemini-3-flash-preview",
-                            "temperature": 0.4,
-                        },
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "Gemini": {
-                        "id": "gemini_default",
                         "provider": "google",
                         "type": "googlegenai_chat_completion",
                         "provider_type": "chat_completion",
@@ -962,10 +884,6 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://generativelanguage.googleapis.com/",
                         "timeout": 120,
-                        "model_config": {
-                            "model": "gemini-3-flash-preview",
-                            "temperature": 0.4,
-                        },
                         "gm_resp_image_modal": False,
                         "gm_native_search": False,
                         "gm_native_coderunner": False,
@@ -977,10 +895,42 @@ CONFIG_METADATA_2 = {
                             "dangerous_content": "BLOCK_MEDIUM_AND_ABOVE",
                         },
                         "gm_thinking_config": {"budget": 0, "level": "HIGH"},
-                        "modalities": ["text", "image", "tool_use"],
+                    },
+                    "Anthropic": {
+                        "id": "anthropic",
+                        "provider": "anthropic",
+                        "type": "anthropic_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "api_base": "https://api.anthropic.com/v1",
+                        "timeout": 120,
+                    },
+                    "Moonshot": {
+                        "id": "moonshot",
+                        "provider": "moonshot",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "timeout": 120,
+                        "api_base": "https://api.moonshot.cn/v1",
+                        "custom_headers": {},
+                    },
+                    "xAI": {
+                        "id": "xai",
+                        "provider": "xai",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "api_base": "https://api.x.ai/v1",
+                        "timeout": 120,
+                        "custom_headers": {},
+                        "xai_native_search": False,
                     },
                     "DeepSeek": {
-                        "id": "deepseek_default",
+                        "id": "deepseek",
                         "provider": "deepseek",
                         "type": "openai_chat_completion",
                         "provider_type": "chat_completion",
@@ -988,13 +938,75 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.deepseek.com/v1",
                         "timeout": 120,
-                        "model_config": {"model": "deepseek-chat", "temperature": 0.4},
                         "custom_headers": {},
-                        "custom_extra_body": {},
+                    },
-                        "modalities": ["text", "tool_use"],
+                    "Zhipu": {
+                        "id": "zhipu",
+                        "provider": "zhipu",
+                        "type": "zhipu_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "timeout": 120,
+                        "api_base": "https://open.bigmodel.cn/api/paas/v4/",
+                        "custom_headers": {},
+                    },
+                    "Azure OpenAI": {
+                        "id": "azure_openai",
+                        "provider": "azure",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "api_version": "2024-05-01-preview",
+                        "key": [],
+                        "api_base": "",
+                        "timeout": 120,
+                        "custom_headers": {},
+                    },
+                    "Ollama": {
+                        "id": "ollama",
+                        "provider": "ollama",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": ["ollama"],  # ollama 的 key 默认是 ollama
+                        "api_base": "http://127.0.0.1:11434/v1",
+                        "custom_headers": {},
+                    },
+                    "LM Studio": {
+                        "id": "lm_studio",
+                        "provider": "lm_studio",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": ["lmstudio"],
+                        "api_base": "http://127.0.0.1:1234/v1",
+                        "custom_headers": {},
+                    },
+                    "ModelStack": {
+                        "id": "modelstack",
+                        "provider": "modelstack",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "api_base": "https://modelstack.app/v1",
+                        "timeout": 120,
+                        "custom_headers": {},
+                    },
+                    "Gemini_OpenAI_API": {
+                        "id": "google_gemini_openai",
+                        "provider": "google",
+                        "type": "openai_chat_completion",
+                        "provider_type": "chat_completion",
+                        "enable": True,
+                        "key": [],
+                        "api_base": "https://generativelanguage.googleapis.com/v1beta/openai/",
+                        "timeout": 120,
+                        "custom_headers": {},
                     },
                     "Groq": {
-                        "id": "groq_default",
+                        "id": "groq",
                         "provider": "groq",
                         "type": "groq_chat_completion",
                         "provider_type": "chat_completion",
@@ -1002,13 +1014,7 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.groq.com/openai/v1",
                         "timeout": 120,
-                        "model_config": {
-                            "model": "openai/gpt-oss-20b",
-                            "temperature": 0.4,
-                        },
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "tool_use"],
                     },
                     "302.AI": {
                         "id": "302ai",
@@ -1019,12 +1025,9 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.302.ai/v1",
                         "timeout": 120,
-                        "model_config": {"model": "gpt-4.1-mini", "temperature": 0.4},
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
                     },
-                    "硅基流动": {
+                    "SiliconFlow": {
                         "id": "siliconflow",
                         "provider": "siliconflow",
                         "type": "openai_chat_completion",
@@ -1033,15 +1036,9 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "timeout": 120,
                         "api_base": "https://api.siliconflow.cn/v1",
-                        "model_config": {
-                            "model": "deepseek-ai/DeepSeek-V3",
-                            "temperature": 0.4,
-                        },
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
                     },
-                    "PPIO派欧云": {
+                    "PPIO": {
                         "id": "ppio",
                         "provider": "ppio",
                         "type": "openai_chat_completion",
@@ -1050,14 +1047,9 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.ppinfra.com/v3/openai",
                         "timeout": 120,
-                        "model_config": {
-                            "model": "deepseek/deepseek-r1",
-                            "temperature": 0.4,
-                        },
                         "custom_headers": {},
-                        "custom_extra_body": {},
                     },
-                    "小马算力": {
+                    "TokenPony": {
                         "id": "tokenpony",
                         "provider": "tokenpony",
                         "type": "openai_chat_completion",
@@ -1066,14 +1058,9 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.tokenpony.cn/v1",
                         "timeout": 120,
-                        "model_config": {
-                            "model": "kimi-k2-instruct-0905",
-                            "temperature": 0.7,
-                        },
                         "custom_headers": {},
-                        "custom_extra_body": {},
                     },
-                    "优云智算": {
+                    "Compshare": {
                         "id": "compshare",
                         "provider": "compshare",
                         "type": "openai_chat_completion",
@@ -1082,42 +1069,18 @@ CONFIG_METADATA_2 = {
                         "key": [],
                         "api_base": "https://api.modelverse.cn/v1",
                         "timeout": 120,
-                        "model_config": {
-                            "model": "moonshotai/Kimi-K2-Instruct",
-                        },
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
                     },
-                    "Kimi": {
+                    "ModelScope": {
-                        "id": "moonshot",
+                        "id": "modelscope",
-                        "provider": "moonshot",
+                        "provider": "modelscope",
                         "type": "openai_chat_completion",
                         "provider_type": "chat_completion",
                         "enable": True,
                         "key": [],
                         "timeout": 120,
-                        "api_base": "https://api.moonshot.cn/v1",
+                        "api_base": "https://api-inference.modelscope.cn/v1",
-                        "model_config": {"model": "moonshot-v1-8k", "temperature": 0.4},
                         "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
-                    "智谱 AI": {
-                        "id": "zhipu_default",
-                        "provider": "zhipu",
-                        "type": "zhipu_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": [],
-                        "timeout": 120,
-                        "api_base": "https://open.bigmodel.cn/api/paas/v4/",
-                        "model_config": {
-                            "model": "glm-4-flash",
-                        },
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
                     },
                     "Dify": {
                         "id": "dify_app_default",
@@ -1132,7 +1095,6 @@ CONFIG_METADATA_2 = {
                         "dify_query_input_key": "astrbot_text_query",
                         "variables": {},
                         "timeout": 60,
-                        "hint": "请确保你在 AstrBot 里设置的 APP 类型和 Dify 里面创建的应用的类型一致！",
                     },
                     "Coze": {
                         "id": "coze",
@@ -1163,20 +1125,6 @@ CONFIG_METADATA_2 = {
                         "variables": {},
                         "timeout": 60,
                     },
-                    "ModelScope": {
-                        "id": "modelscope",
-                        "provider": "modelscope",
-                        "type": "openai_chat_completion",
-                        "provider_type": "chat_completion",
-                        "enable": True,
-                        "key": [],
-                        "timeout": 120,
-                        "api_base": "https://api-inference.modelscope.cn/v1",
-                        "model_config": {"model": "Qwen/Qwen3-32B", "temperature": 0.4},
-                        "custom_headers": {},
-                        "custom_extra_body": {},
-                        "modalities": ["text", "image", "tool_use"],
-                    },
                     "FastGPT": {
                         "id": "fastgpt",
                         "provider": "fastgpt",
@@ -1200,7 +1148,6 @@ CONFIG_METADATA_2 = {
                         "model": "whisper-1",
                     },
                     "Whisper(Local)": {
-                        "hint": "启用前请 pip 安装 openai-whisper 库（N卡用户大约下载 2GB，主要是 torch 和 cuda，CPU 用户大约下载 1 GB），并且安装 ffmpeg。否则将无法正常转文字。",
                         "provider": "openai",
                         "type": "openai_whisper_selfhost",
                         "provider_type": "speech_to_text",
@@ -1209,7 +1156,6 @@ CONFIG_METADATA_2 = {
                         "model": "tiny",
                     },
                     "SenseVoice(Local)": {
-                        "hint": "启用前请 pip 安装 funasr、funasr_onnx、torchaudio、torch、modelscope、jieba 库（默认使用CPU，大约下载 1 GB），并且安装 ffmpeg。否则将无法正常转文字。",
                         "type": "sensevoice_stt_selfhost",
                         "provider": "sensevoice",
                         "provider_type": "speech_to_text",
@@ -1231,7 +1177,6 @@ CONFIG_METADATA_2 = {
                         "timeout": "20",
                     },
                     "Edge TTS": {
-                        "hint": "提示：使用这个服务前需要安装有 ffmpeg，并且可以直接在终端调用 ffmpeg 指令。",
                         "id": "edge_tts",
                         "provider": "microsoft",
                         "type": "edge_tts",
@@ -1447,6 +1392,10 @@ CONFIG_METADATA_2 = {
                     },
                 },
                 "items": {
+                    "provider_source_id": {
+                        "invisible": True,
+                        "type": "string",
+                    },
                     "xai_native_search": {
                         "description": "启用原生搜索功能",
                         "type": "bool",
@@ -2015,7 +1964,6 @@ CONFIG_METADATA_2 = {
                     "id": {
                         "description": "ID",
                         "type": "string",
-                        "hint": "模型提供商名字。",
                     },
                     "type": {
                         "description": "模型提供商种类",
@@ -2035,29 +1983,15 @@ CONFIG_METADATA_2 = {
                         "description": "API Key",
                         "type": "list",
                         "items": {"type": "string"},
-                        "hint": "提供商 API Key。",
                     },
                     "api_base": {
                         "description": "API Base URL",
                         "type": "string",
-                        "hint": "API Base URL 请在模型提供商处获得。如出现 404 报错，尝试在地址末尾加上 /v1",
                     },
-                    "model_config": {
+                    "model": {
-                        "description": "模型配置",
+                        "description": "模型 ID",
-                        "type": "object",
+                        "type": "string",
-                        "items": {
+                        "hint": "模型名称，如 gpt-4o-mini, deepseek-chat。",
-                            "model": {
-                                "description": "模型名称",
-                                "type": "string",
-                                "hint": "模型名称，如 gpt-4o-mini, deepseek-chat。",
-                            },
-                            "max_tokens": {
-                                "description": "模型最大输出长度（tokens）",
-                                "type": "int",
-                            },
-                            "temperature": {"description": "温度", "type": "float"},
-                            "top_p": {"description": "Top P值", "type": "float"},
-                        },
                     },
                     "dify_api_key": {
                         "description": "API Key",

astrbot/core/core_lifecycle.py

@@ -33,6 +33,7 @@ from astrbot.core.star.context import Context
 from astrbot.core.star.star_handler import EventType, star_handlers_registry, star_map
 from astrbot.core.umop_config_router import UmopConfigRouter
 from astrbot.core.updator import AstrBotUpdator
+from astrbot.core.utils.llm_metadata import update_llm_metadata
 from astrbot.core.utils.migra_helper import migra
 
 from . import astrbot_config, html_renderer
@@ -185,6 +186,8 @@ class AstrBotCoreLifecycle:
         # 初始化关闭控制面板的事件
         self.dashboard_shutdown_event = asyncio.Event()
 
+        asyncio.create_task(update_llm_metadata())
+
     def _load(self) -> None:
         """加载事件总线和任务并初始化."""
         # 创建一个异步任务来执行事件总线的 dispatch() 方法

astrbot/core/pipeline/process_stage/method/agent_sub_stages/internal.py

@@ -321,7 +321,12 @@ class InternalAgentSubStage(Stage):
             elif isinstance(req.tool_calls_result, list):
                 for tcr in req.tool_calls_result:
                     messages.extend(tcr.to_openai_messages())
-        messages.append({"role": "assistant", "content": llm_response.completion_text})
+        messages.append(
+            {
+                "role": "assistant",
+                "content": llm_response.completion_text or "*No response*",
+            }
+        )
         messages = list(filter(lambda item: "_no_save" not in item, messages))
         await self.conv_manager.update_conversation(
             event.unified_msg_origin,

astrbot/core/platform/sources/aiocqhttp/aiocqhttp_platform_adapter.py

@@ -385,10 +385,25 @@ class AiocqhttpAdapter(Platform):
                         logger.error(f"获取 @ 用户信息失败: {e}，此消息段将被忽略。")
 
                 message_str += "".join(at_parts)
+            elif t == "markdown":
+                text = m["data"].get("markdown") or m["data"].get("content", "")
+                abm.message.append(Plain(text=text))
+                message_str += text
             else:
                 for m in m_group:
-                    a = ComponentTypes[t](**m["data"])
+                    try:
-                    abm.message.append(a)
+                        if t not in ComponentTypes:
+                            logger.warning(
+                                f"不支持的消息段类型，已忽略: {t}, data={m['data']}"
+                            )
+                            continue
+                        a = ComponentTypes[t](**m["data"])
+                        abm.message.append(a)
+                    except Exception as e:
+                        logger.exception(
+                            f"消息段解析失败: type={t}, data={m['data']}. {e}"
+                        )
+                        continue
 
         abm.timestamp = int(time.time())
         abm.message_str = message_str

astrbot/core/provider/entities.py

@@ -14,6 +14,7 @@ import astrbot.core.message.components as Comp
 from astrbot import logger
 from astrbot.core.agent.message import (
     AssistantMessageSegment,
+    ContentPart,
     ToolCall,
     ToolCallMessageSegment,
 )
@@ -92,6 +93,8 @@ class ProviderRequest:
     """会话 ID"""
     image_urls: list[str] = field(default_factory=list)
     """图片 URL 列表"""
+    extra_user_content_parts: list[ContentPart] = field(default_factory=list)
+    """额外的用户消息内容部分列表，用于在用户消息后添加额外的内容块（如系统提醒、指令等）。"""
     func_tool: ToolSet | None = None
     """可用的函数工具"""
     contexts: list[dict] = field(default_factory=list)
@@ -166,13 +169,23 @@ class ProviderRequest:
 
     async def assemble_context(self) -> dict:
         """将请求(prompt 和 image_urls)包装成 OpenAI 的消息格式。"""
+        # 构建内容块列表
+        content_blocks = []
+
+        # 1. 用户原始发言（OpenAI 建议：用户发言在前）
+        if self.prompt and self.prompt.strip():
+            content_blocks.append({"type": "text", "text": self.prompt})
+        elif self.image_urls:
+            # 如果没有文本但有图片，添加占位文本
+            content_blocks.append({"type": "text", "text": "[图片]"})
+
+        # 2. 额外的内容块（系统提醒、指令等）
+        if self.extra_user_content_parts:
+            for part in self.extra_user_content_parts:
+                content_blocks.append(part.model_dump())
+
+        # 3. 图片内容
         if self.image_urls:
-            user_content = {
-                "role": "user",
-                "content": [
-                    {"type": "text", "text": self.prompt if self.prompt else "[图片]"},
-                ],
-            }
             for image_url in self.image_urls:
                 if image_url.startswith("http"):
                     image_path = await download_image_by_url(image_url)
@@ -185,11 +198,21 @@ class ProviderRequest:
                 if not image_data:
                     logger.warning(f"图片 {image_url} 得到的结果为空，将忽略。")
                     continue
-                user_content["content"].append(
+                content_blocks.append(
                     {"type": "image_url", "image_url": {"url": image_data}},
                 )
-            return user_content
+
-        return {"role": "user", "content": self.prompt}
+        # 只有当只有一个来自 prompt 的文本块且没有额外内容块时，才降级为简单格式以保持向后兼容
+        if (
+            len(content_blocks) == 1
+            and content_blocks[0]["type"] == "text"
+            and not self.extra_user_content_parts
+            and not self.image_urls
+        ):
+            return {"role": "user", "content": content_blocks[0]["text"]}
+
+        # 否则返回多模态格式
+        return {"role": "user", "content": content_blocks}
 
     async def _encode_image_bs64(self, image_url: str) -> str:
         """将图片转换为 base64"""

astrbot/core/provider/manager.py

@@ -1,4 +1,5 @@
 import asyncio
+import copy
 import traceback
 from typing import Protocol, runtime_checkable
 
@@ -32,10 +33,12 @@ class ProviderManager:
         persona_mgr: PersonaManager,
     ):
         self.reload_lock = asyncio.Lock()
+        self.resource_lock = asyncio.Lock()
         self.persona_mgr = persona_mgr
         self.acm = acm
         config = acm.confs["default"]
         self.providers_config: list = config["provider"]
+        self.provider_sources_config: list = config.get("provider_sources", [])
         self.provider_settings: dict = config["provider_settings"]
         self.provider_stt_settings: dict = config.get("provider_stt_settings", {})
         self.provider_tts_settings: dict = config.get("provider_tts_settings", {})
@@ -148,6 +151,7 @@ class ProviderManager:
 
         """
         provider = None
+        provider_id = None
         if umo:
             provider_id = sp.get(
                 f"provider_perf_{provider_type.value}",
@@ -185,6 +189,12 @@ class ProviderManager:
                     )
             else:
                 raise ValueError(f"Unknown provider type: {provider_type}")
+
+        if not provider and provider_id:
+            logger.warning(
+                f"没有找到 ID 为 {provider_id} 的提供商，这可能是由于您修改了提供商（模型）ID 导致的。"
+            )
+
         return provider
 
     async def initialize(self):
@@ -251,7 +261,136 @@ class ProviderManager:
         # 初始化 MCP Client 连接
         asyncio.create_task(self.llm_tools.init_mcp_clients(), name="init_mcp_clients")
 
+    def dynamic_import_provider(self, type: str):
+        """动态导入提供商适配器模块
+
+        Args:
+            type (str): 提供商请求类型。
+
+        Raises:
+            ImportError: 如果提供商类型未知或无法导入对应模块，则抛出异常。
+        """
+        match type:
+            case "openai_chat_completion":
+                from .sources.openai_source import (
+                    ProviderOpenAIOfficial as ProviderOpenAIOfficial,
+                )
+            case "zhipu_chat_completion":
+                from .sources.zhipu_source import ProviderZhipu as ProviderZhipu
+            case "groq_chat_completion":
+                from .sources.groq_source import ProviderGroq as ProviderGroq
+            case "anthropic_chat_completion":
+                from .sources.anthropic_source import (
+                    ProviderAnthropic as ProviderAnthropic,
+                )
+            case "googlegenai_chat_completion":
+                from .sources.gemini_source import (
+                    ProviderGoogleGenAI as ProviderGoogleGenAI,
+                )
+            case "sensevoice_stt_selfhost":
+                from .sources.sensevoice_selfhosted_source import (
+                    ProviderSenseVoiceSTTSelfHost as ProviderSenseVoiceSTTSelfHost,
+                )
+            case "openai_whisper_api":
+                from .sources.whisper_api_source import (
+                    ProviderOpenAIWhisperAPI as ProviderOpenAIWhisperAPI,
+                )
+            case "openai_whisper_selfhost":
+                from .sources.whisper_selfhosted_source import (
+                    ProviderOpenAIWhisperSelfHost as ProviderOpenAIWhisperSelfHost,
+                )
+            case "xinference_stt":
+                from .sources.xinference_stt_provider import (
+                    ProviderXinferenceSTT as ProviderXinferenceSTT,
+                )
+            case "openai_tts_api":
+                from .sources.openai_tts_api_source import (
+                    ProviderOpenAITTSAPI as ProviderOpenAITTSAPI,
+                )
+            case "edge_tts":
+                from .sources.edge_tts_source import (
+                    ProviderEdgeTTS as ProviderEdgeTTS,
+                )
+            case "gsv_tts_selfhost":
+                from .sources.gsv_selfhosted_source import (
+                    ProviderGSVTTS as ProviderGSVTTS,
+                )
+            case "gsvi_tts_api":
+                from .sources.gsvi_tts_source import (
+                    ProviderGSVITTS as ProviderGSVITTS,
+                )
+            case "fishaudio_tts_api":
+                from .sources.fishaudio_tts_api_source import (
+                    ProviderFishAudioTTSAPI as ProviderFishAudioTTSAPI,
+                )
+            case "dashscope_tts":
+                from .sources.dashscope_tts import (
+                    ProviderDashscopeTTSAPI as ProviderDashscopeTTSAPI,
+                )
+            case "azure_tts":
+                from .sources.azure_tts_source import (
+                    AzureTTSProvider as AzureTTSProvider,
+                )
+            case "minimax_tts_api":
+                from .sources.minimax_tts_api_source import (
+                    ProviderMiniMaxTTSAPI as ProviderMiniMaxTTSAPI,
+                )
+            case "volcengine_tts":
+                from .sources.volcengine_tts import (
+                    ProviderVolcengineTTS as ProviderVolcengineTTS,
+                )
+            case "gemini_tts":
+                from .sources.gemini_tts_source import (
+                    ProviderGeminiTTSAPI as ProviderGeminiTTSAPI,
+                )
+            case "openai_embedding":
+                from .sources.openai_embedding_source import (
+                    OpenAIEmbeddingProvider as OpenAIEmbeddingProvider,
+                )
+            case "gemini_embedding":
+                from .sources.gemini_embedding_source import (
+                    GeminiEmbeddingProvider as GeminiEmbeddingProvider,
+                )
+            case "vllm_rerank":
+                from .sources.vllm_rerank_source import (
+                    VLLMRerankProvider as VLLMRerankProvider,
+                )
+            case "xinference_rerank":
+                from .sources.xinference_rerank_source import (
+                    XinferenceRerankProvider as XinferenceRerankProvider,
+                )
+            case "bailian_rerank":
+                from .sources.bailian_rerank_source import (
+                    BailianRerankProvider as BailianRerankProvider,
+                )
+
+    def get_merged_provider_config(self, provider_config: dict) -> dict:
+        """获取 provider 配置和 provider_source 配置合并后的结果
+
+        Returns:
+            dict: 合并后的 provider 配置，key 为 provider id，value 为合并后的配置字典
+        """
+        pc = copy.deepcopy(provider_config)
+        provider_source_id = pc.get("provider_source_id", "")
+        if provider_source_id:
+            provider_source = None
+            for ps in self.provider_sources_config:
+                if ps.get("id") == provider_source_id:
+                    provider_source = ps
+                    break
+
+            if provider_source:
+                # 合并配置，provider 的配置优先级更高
+                merged_config = {**provider_source, **pc}
+                # 保持 id 为 provider 的 id，而不是 source 的 id
+                merged_config["id"] = pc["id"]
+                pc = merged_config
+        return pc
+
     async def load_provider(self, provider_config: dict):
+        # 如果 provider_source_id 存在且不为空，则从 provider_sources 中找到对应的配置并合并
+        provider_config = self.get_merged_provider_config(provider_config)
+
         if not provider_config["enable"]:
             logger.info(f"Provider {provider_config['id']} is disabled, skipping")
             return
@@ -264,99 +403,7 @@ class ProviderManager:
 
         # 动态导入
         try:
-            match provider_config["type"]:
+            self.dynamic_import_provider(provider_config["type"])
-                case "openai_chat_completion":
-                    from .sources.openai_source import (
-                        ProviderOpenAIOfficial as ProviderOpenAIOfficial,
-                    )
-                case "zhipu_chat_completion":
-                    from .sources.zhipu_source import ProviderZhipu as ProviderZhipu
-                case "groq_chat_completion":
-                    from .sources.groq_source import ProviderGroq as ProviderGroq
-                case "anthropic_chat_completion":
-                    from .sources.anthropic_source import (
-                        ProviderAnthropic as ProviderAnthropic,
-                    )
-                case "googlegenai_chat_completion":
-                    from .sources.gemini_source import (
-                        ProviderGoogleGenAI as ProviderGoogleGenAI,
-                    )
-                case "sensevoice_stt_selfhost":
-                    from .sources.sensevoice_selfhosted_source import (
-                        ProviderSenseVoiceSTTSelfHost as ProviderSenseVoiceSTTSelfHost,
-                    )
-                case "openai_whisper_api":
-                    from .sources.whisper_api_source import (
-                        ProviderOpenAIWhisperAPI as ProviderOpenAIWhisperAPI,
-                    )
-                case "openai_whisper_selfhost":
-                    from .sources.whisper_selfhosted_source import (
-                        ProviderOpenAIWhisperSelfHost as ProviderOpenAIWhisperSelfHost,
-                    )
-                case "xinference_stt":
-                    from .sources.xinference_stt_provider import (
-                        ProviderXinferenceSTT as ProviderXinferenceSTT,
-                    )
-                case "openai_tts_api":
-                    from .sources.openai_tts_api_source import (
-                        ProviderOpenAITTSAPI as ProviderOpenAITTSAPI,
-                    )
-                case "edge_tts":
-                    from .sources.edge_tts_source import (
-                        ProviderEdgeTTS as ProviderEdgeTTS,
-                    )
-                case "gsv_tts_selfhost":
-                    from .sources.gsv_selfhosted_source import (
-                        ProviderGSVTTS as ProviderGSVTTS,
-                    )
-                case "gsvi_tts_api":
-                    from .sources.gsvi_tts_source import (
-                        ProviderGSVITTS as ProviderGSVITTS,
-                    )
-                case "fishaudio_tts_api":
-                    from .sources.fishaudio_tts_api_source import (
-                        ProviderFishAudioTTSAPI as ProviderFishAudioTTSAPI,
-                    )
-                case "dashscope_tts":
-                    from .sources.dashscope_tts import (
-                        ProviderDashscopeTTSAPI as ProviderDashscopeTTSAPI,
-                    )
-                case "azure_tts":
-                    from .sources.azure_tts_source import (
-                        AzureTTSProvider as AzureTTSProvider,
-                    )
-                case "minimax_tts_api":
-                    from .sources.minimax_tts_api_source import (
-                        ProviderMiniMaxTTSAPI as ProviderMiniMaxTTSAPI,
-                    )
-                case "volcengine_tts":
-                    from .sources.volcengine_tts import (
-                        ProviderVolcengineTTS as ProviderVolcengineTTS,
-                    )
-                case "gemini_tts":
-                    from .sources.gemini_tts_source import (
-                        ProviderGeminiTTSAPI as ProviderGeminiTTSAPI,
-                    )
-                case "openai_embedding":
-                    from .sources.openai_embedding_source import (
-                        OpenAIEmbeddingProvider as OpenAIEmbeddingProvider,
-                    )
-                case "gemini_embedding":
-                    from .sources.gemini_embedding_source import (
-                        GeminiEmbeddingProvider as GeminiEmbeddingProvider,
-                    )
-                case "vllm_rerank":
-                    from .sources.vllm_rerank_source import (
-                        VLLMRerankProvider as VLLMRerankProvider,
-                    )
-                case "xinference_rerank":
-                    from .sources.xinference_rerank_source import (
-                        XinferenceRerankProvider as XinferenceRerankProvider,
-                    )
-                case "bailian_rerank":
-                    from .sources.bailian_rerank_source import (
-                        BailianRerankProvider as BailianRerankProvider,
-                    )
         except (ImportError, ModuleNotFoundError) as e:
             logger.critical(
                 f"加载 {provider_config['type']}({provider_config['id']}) 提供商适配器失败：{e}。可能是因为有未安装的依赖。",
@@ -499,6 +546,7 @@ class ProviderManager:
 
             # 和配置文件保持同步
             self.providers_config = astrbot_config["provider"]
+            self.provider_sources_config = astrbot_config.get("provider_sources", [])
             config_ids = [provider["id"] for provider in self.providers_config]
             logger.info(f"providers in user's config: {config_ids}")
             for key in list(self.inst_map.keys()):
@@ -570,6 +618,68 @@ class ProviderManager:
             )
             del self.inst_map[provider_id]
 
+    async def delete_provider(
+        self, provider_id: str | None = None, provider_source_id: str | None = None
+    ):
+        """Delete provider and/or provider source from config and terminate the instances. Config will be saved after deletion."""
+        async with self.resource_lock:
+            # delete from config
+            target_prov_ids = []
+            if provider_id:
+                target_prov_ids.append(provider_id)
+            else:
+                for prov in self.providers_config:
+                    if prov.get("provider_source_id") == provider_source_id:
+                        target_prov_ids.append(prov.get("id"))
+            config = self.acm.default_conf
+            for tpid in target_prov_ids:
+                await self.terminate_provider(tpid)
+                config["provider"] = [
+                    prov for prov in config["provider"] if prov.get("id") != tpid
+                ]
+            config.save_config()
+            logger.info(f"Provider {target_prov_ids} 已从配置中删除。")
+
+    async def update_provider(self, origin_provider_id: str, new_config: dict):
+        """Update provider config and reload the instance. Config will be saved after update."""
+        async with self.resource_lock:
+            npid = new_config.get("id", None)
+            if not npid:
+                raise ValueError("New provider config must have an 'id' field")
+            config = self.acm.default_conf
+            for provider in config["provider"]:
+                if (
+                    provider.get("id", None) == npid
+                    and provider.get("id", None) != origin_provider_id
+                ):
+                    raise ValueError(f"Provider ID {npid} already exists")
+            # update config
+            for idx, provider in enumerate(config["provider"]):
+                if provider.get("id", None) == origin_provider_id:
+                    config["provider"][idx] = new_config
+                    break
+            else:
+                raise ValueError(f"Provider ID {origin_provider_id} not found")
+            config.save_config()
+            # reload instance
+            await self.reload(new_config)
+
+    async def create_provider(self, new_config: dict):
+        """Add new provider config and load the instance. Config will be saved after addition."""
+        async with self.resource_lock:
+            npid = new_config.get("id", None)
+            if not npid:
+                raise ValueError("New provider config must have an 'id' field")
+            config = self.acm.default_conf
+            for provider in config["provider"]:
+                if provider.get("id", None) == npid:
+                    raise ValueError(f"Provider ID {npid} already exists")
+            # add to config
+            config["provider"].append(new_config)
+            config.save_config()
+            # load instance
+            await self.load_provider(new_config)
+
     async def terminate(self):
         for provider_inst in self.provider_insts:
             if hasattr(provider_inst, "terminate"):

astrbot/core/provider/provider.py

@@ -4,7 +4,7 @@ import os
 from collections.abc import AsyncGenerator
 from typing import TypeAlias, Union
 
-from astrbot.core.agent.message import Message
+from astrbot.core.agent.message import ContentPart, Message
 from astrbot.core.agent.tool import ToolSet
 from astrbot.core.provider.entities import (
     LLMResponse,
@@ -103,6 +103,7 @@ class Provider(AbstractProvider):
         system_prompt: str | None = None,
         tool_calls_result: ToolCallsResult | list[ToolCallsResult] | None = None,
         model: str | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
         **kwargs,
     ) -> LLMResponse:
         """获得 LLM 的文本对话结果。会使用当前的模型进行对话。
@@ -114,6 +115,7 @@ class Provider(AbstractProvider):
             tools: tool set
             contexts: 上下文，和 prompt 二选一使用
             tool_calls_result: 回传给 LLM 的工具调用结果。参考: https://platform.openai.com/docs/guides/function-calling
+            extra_user_content_parts: 额外的用户内容块列表，用于在用户消息后添加额外的文本块（如系统提醒、指令等）
             kwargs: 其他参数
 
         Notes:
@@ -133,6 +135,7 @@ class Provider(AbstractProvider):
         system_prompt: str | None = None,
         tool_calls_result: ToolCallsResult | list[ToolCallsResult] | None = None,
         model: str | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
         **kwargs,
     ) -> AsyncGenerator[LLMResponse, None]:
         """获得 LLM 的流式文本对话结果。会使用当前的模型进行对话。在生成的最后会返回一次完整的结果。
@@ -144,6 +147,7 @@ class Provider(AbstractProvider):
             tools: tool set
             contexts: 上下文，和 prompt 二选一使用
             tool_calls_result: 回传给 LLM 的工具调用结果。参考: https://platform.openai.com/docs/guides/function-calling
+            extra_user_content_parts: 额外的用户内容块列表，用于在用户消息后添加额外的文本块（如系统提醒、指令等）
             kwargs: 其他参数
 
         Notes:

astrbot/core/provider/sources/anthropic_source.py

@@ -11,6 +11,7 @@ from anthropic.types.usage import Usage
 
 from astrbot import logger
 from astrbot.api.provider import Provider
+from astrbot.core.agent.message import ContentPart
 from astrbot.core.provider.entities import LLMResponse, TokenUsage
 from astrbot.core.provider.func_tool_manager import ToolSet
 from astrbot.core.utils.io import download_image_by_url
@@ -47,7 +48,7 @@ class ProviderAnthropic(Provider):
             base_url=self.base_url,
         )
 
-        self.set_model(provider_config["model_config"]["model"])
+        self.set_model(provider_config.get("model", "unknown"))
 
     def _prepare_payload(self, messages: list[dict]):
         """准备 Anthropic API 的请求 payload
@@ -130,7 +131,11 @@ class ProviderAnthropic(Provider):
             if tool_list := tools.get_func_desc_anthropic_style():
                 payloads["tools"] = tool_list
 
-        completion = await self.client.messages.create(**payloads, stream=False)
+        extra_body = self.provider_config.get("custom_extra_body", {})
+
+        completion = await self.client.messages.create(
+            **payloads, stream=False, extra_body=extra_body
+        )
 
         assert isinstance(completion, Message)
         logger.debug(f"completion: {completion}")
@@ -173,11 +178,13 @@ class ProviderAnthropic(Provider):
         # 用于累积最终结果
         final_text = ""
         final_tool_calls = []
-
         id = None
         usage = TokenUsage()
+        extra_body = self.provider_config.get("custom_extra_body", {})
 
-        async with self.client.messages.stream(**payloads) as stream:
+        async with self.client.messages.stream(
+            **payloads, extra_body=extra_body
+        ) as stream:
             assert isinstance(stream, anthropic.AsyncMessageStream)
             async for event in stream:
                 if event.type == "message_start":
@@ -290,13 +297,16 @@ class ProviderAnthropic(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ) -> LLMResponse:
         if contexts is None:
             contexts = []
         new_record = None
         if prompt is not None:
-            new_record = await self.assemble_context(prompt, image_urls)
+            new_record = await self.assemble_context(
+                prompt, image_urls, extra_user_content_parts
+            )
         context_query = self._ensure_message_to_dicts(contexts)
         if new_record:
             context_query.append(new_record)
@@ -318,10 +328,9 @@ class ProviderAnthropic(Provider):
 
         system_prompt, new_messages = self._prepare_payload(context_query)
 
-        model_config = self.provider_config.get("model_config", {})
+        model = model or self.get_model()
-        model_config["model"] = model or self.get_model()
 
-        payloads = {"messages": new_messages, **model_config}
+        payloads = {"messages": new_messages, "model": model}
 
         # Anthropic has a different way of handling system prompts
         if system_prompt:
@@ -331,7 +340,6 @@ class ProviderAnthropic(Provider):
         try:
             llm_response = await self._query(payloads, func_tool)
         except Exception as e:
-            # logger.error(f"发生了错误。Provider 配置如下: {model_config}")
             raise e
 
         return llm_response
@@ -346,13 +354,16 @@ class ProviderAnthropic(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ):
         if contexts is None:
             contexts = []
         new_record = None
         if prompt is not None:
-            new_record = await self.assemble_context(prompt, image_urls)
+            new_record = await self.assemble_context(
+                prompt, image_urls, extra_user_content_parts
+            )
         context_query = self._ensure_message_to_dicts(contexts)
         if new_record:
             context_query.append(new_record)
@@ -373,10 +384,9 @@ class ProviderAnthropic(Provider):
 
         system_prompt, new_messages = self._prepare_payload(context_query)
 
-        model_config = self.provider_config.get("model_config", {})
+        model = model or self.get_model()
-        model_config["model"] = model or self.get_model()
 
-        payloads = {"messages": new_messages, **model_config}
+        payloads = {"messages": new_messages, "model": model}
 
         # Anthropic has a different way of handling system prompts
         if system_prompt:
@@ -385,48 +395,116 @@ class ProviderAnthropic(Provider):
         async for llm_response in self._query_stream(payloads, func_tool):
             yield llm_response
 
-    async def assemble_context(self, text: str, image_urls: list[str] | None = None):
+    async def assemble_context(
+        self,
+        text: str,
+        image_urls: list[str] | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
+    ):
         """组装上下文，支持文本和图片"""
-        if not image_urls:
-            return {"role": "user", "content": text}
-
         content = []
-        content.append({"type": "text", "text": text})
 
-        for image_url in image_urls:
+        # 1. 用户原始发言（OpenAI 建议：用户发言在前）
-            if image_url.startswith("http"):
+        if text:
-                image_path = await download_image_by_url(image_url)
+            content.append({"type": "text", "text": text})
-                image_data = await self.encode_image_bs64(image_path)
+        elif image_urls:
-            elif image_url.startswith("file:///"):
+            # 如果没有文本但有图片，添加占位文本
-                image_path = image_url.replace("file:///", "")
+            content.append({"type": "text", "text": "[图片]"})
-                image_data = await self.encode_image_bs64(image_path)
+        elif extra_user_content_parts:
-            else:
+            # 如果只有额外内容块，也需要添加占位文本
-                image_data = await self.encode_image_bs64(image_url)
+            content.append({"type": "text", "text": " "})
 
-            if not image_data:
+        # 2. 额外的内容块（系统提醒、指令等）
-                logger.warning(f"图片 {image_url} 得到的结果为空，将忽略。")
+        if extra_user_content_parts:
-                continue
+            for block in extra_user_content_parts:
+                block_type = block.get("type")
 
-            # Get mime type for the image
+                if block_type == "text":
-            mime_type, _ = guess_type(image_url)
+                    # 文本直接添加
-            if not mime_type:
+                    content.append(block)
-                mime_type = "image/jpeg"  # Default to JPEG if can't determine
 
-            content.append(
+                elif block_type == "image_url":
-                {
+                    # 转换 OpenAI 格式的图片为 Anthropic 格式
-                    "type": "image",
+                    image_url_data = block.get("image_url", {})
-                    "source": {
+                    if isinstance(image_url_data, dict):
-                        "type": "base64",
+                        url = image_url_data.get("url", "")
-                        "media_type": mime_type,
+                    else:
-                        "data": (
+                        # 兼容直接传 URL 字符串的情况
-                            image_data.split("base64,")[1]
+                        url = str(image_url_data)
-                            if "base64," in image_data
+
-                            else image_data
+                    if url and url.startswith("data:"):
-                        ),
+                        try:
+                            # 提取 MIME 类型和 base64 数据
+                            mime_type = url.split(":")[1].split(";")[0]
+                            base64_data = (
+                                url.split("base64,")[1] if "base64," in url else url
+                            )
+                            content.append(
+                                {
+                                    "type": "image",
+                                    "source": {
+                                        "type": "base64",
+                                        "media_type": mime_type,
+                                        "data": base64_data,
+                                    },
+                                }
+                            )
+                        except Exception as e:
+                            logger.warning(f"转换 image_url 到 Anthropic 格式失败: {e}")
+                    else:
+                        logger.warning(f"image_url 不是有效的 data URI: {url[:50]}...")
+
+                else:
+                    # 其他类型（如 audio_url）Anthropic 不支持，记录警告
+                    logger.debug(f"Anthropic 不支持的内容类型 '{block_type}'，已忽略")
+
+        # 3. 图片内容
+        if image_urls:
+            for image_url in image_urls:
+                if image_url.startswith("http"):
+                    image_path = await download_image_by_url(image_url)
+                    image_data = await self.encode_image_bs64(image_path)
+                elif image_url.startswith("file:///"):
+                    image_path = image_url.replace("file:///", "")
+                    image_data = await self.encode_image_bs64(image_path)
+                else:
+                    image_data = await self.encode_image_bs64(image_url)
+
+                if not image_data:
+                    logger.warning(f"图片 {image_url} 得到的结果为空，将忽略。")
+                    continue
+
+                # Get mime type for the image
+                mime_type, _ = guess_type(image_url)
+                if not mime_type:
+                    mime_type = "image/jpeg"  # Default to JPEG if can't determine
+
+                content.append(
+                    {
+                        "type": "image",
+                        "source": {
+                            "type": "base64",
+                            "media_type": mime_type,
+                            "data": (
+                                image_data.split("base64,")[1]
+                                if "base64," in image_data
+                                else image_data
+                            ),
+                        },
                     },
-                },
+                )
-            )
 
+        # 如果只有主文本且没有额外内容块和图片，返回简单格式以保持向后兼容
+        if (
+            text
+            and not extra_user_content_parts
+            and not image_urls
+            and len(content) == 1
+            and content[0]["type"] == "text"
+        ):
+            return {"role": "user", "content": content[0]["text"]}
+
+        # 否则返回多模态格式
         return {"role": "user", "content": content}
 
     async def encode_image_bs64(self, image_url: str) -> str:

astrbot/core/provider/sources/gemini_source.py

@@ -13,6 +13,7 @@ from google.genai.errors import APIError
 import astrbot.core.message.components as Comp
 from astrbot import logger
 from astrbot.api.provider import Provider
+from astrbot.core.agent.message import ContentPart
 from astrbot.core.message.message_event_result import MessageChain
 from astrbot.core.provider.entities import LLMResponse, TokenUsage
 from astrbot.core.provider.func_tool_manager import ToolSet
@@ -68,7 +69,7 @@ class ProviderGoogleGenAI(Provider):
             self.api_base = self.api_base[:-1]
 
         self._init_client()
-        self.set_model(provider_config["model_config"]["model"])
+        self.set_model(provider_config.get("model", "unknown"))
         self._init_safety_settings()
 
     def _init_client(self) -> None:
@@ -138,7 +139,7 @@ class ProviderGoogleGenAI(Provider):
             modalities = ["TEXT"]
 
         tool_list: list[types.Tool] | None = []
-        model_name = payloads.get("model", self.get_model())
+        model_name = cast(str, payloads.get("model", self.get_model()))
         native_coderunner = self.provider_config.get("gm_native_coderunner", False)
         native_search = self.provider_config.get("gm_native_search", False)
         url_context = self.provider_config.get("gm_url_context", False)
@@ -199,7 +200,16 @@ class ProviderGoogleGenAI(Provider):
 
         # oper thinking config
         thinking_config = None
-        if model_name.startswith("gemini-2.5"):
+        if model_name in [
+            "gemini-2.5-pro",
+            "gemini-2.5-pro-preview",
+            "gemini-2.5-flash",
+            "gemini-2.5-flash-preview",
+            "gemini-2.5-flash-lite",
+            "gemini-2.5-flash-lite-preview",
+            "gemini-robotics-er-1.5-preview",
+            "gemini-live-2.5-flash-preview-native-audio-09-2025",
+        ]:
             # The thinkingBudget parameter, introduced with the Gemini 2.5 series
             thinking_budget = self.provider_config.get("gm_thinking_config", {}).get(
                 "budget", 0
@@ -208,7 +218,14 @@ class ProviderGoogleGenAI(Provider):
                 thinking_config = types.ThinkingConfig(
                     thinking_budget=thinking_budget,
                 )
-        elif model_name.startswith("gemini-3"):
+        elif model_name in [
+            "gemini-3-pro",
+            "gemini-3-pro-preview",
+            "gemini-3-flash",
+            "gemini-3-flash-preview",
+            "gemini-3-flash-lite",
+            "gemini-3-flash-lite-preview",
+        ]:
             # The thinkingLevel parameter, recommended for Gemini 3 models and onwards
             # Gemini 2.5 series models don't support thinkingLevel; use thinkingBudget instead.
             thinking_level = self.provider_config.get("gm_thinking_config", {}).get(
@@ -664,13 +681,16 @@ class ProviderGoogleGenAI(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ) -> LLMResponse:
         if contexts is None:
             contexts = []
         new_record = None
         if prompt is not None:
-            new_record = await self.assemble_context(prompt, image_urls)
+            new_record = await self.assemble_context(
+                prompt, image_urls, extra_user_content_parts
+            )
         context_query = self._ensure_message_to_dicts(contexts)
         if new_record:
             context_query.append(new_record)
@@ -689,10 +709,9 @@ class ProviderGoogleGenAI(Provider):
                 for tcr in tool_calls_result:
                     context_query.extend(tcr.to_openai_messages())
 
-        model_config = self.provider_config.get("model_config", {})
+        model = model or self.get_model()
-        model_config["model"] = model or self.get_model()
 
-        payloads = {"messages": context_query, **model_config}
+        payloads = {"messages": context_query, "model": model}
 
         retry = 10
         keys = self.api_keys.copy()
@@ -717,13 +736,16 @@ class ProviderGoogleGenAI(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ) -> AsyncGenerator[LLMResponse, None]:
         if contexts is None:
             contexts = []
         new_record = None
         if prompt is not None:
-            new_record = await self.assemble_context(prompt, image_urls)
+            new_record = await self.assemble_context(
+                prompt, image_urls, extra_user_content_parts
+            )
         context_query = self._ensure_message_to_dicts(contexts)
         if new_record:
             context_query.append(new_record)
@@ -742,10 +764,9 @@ class ProviderGoogleGenAI(Provider):
                 for tcr in tool_calls_result:
                     context_query.extend(tcr.to_openai_messages())
 
-        model_config = self.provider_config.get("model_config", {})
+        model = model or self.get_model()
-        model_config["model"] = model or self.get_model()
 
-        payloads = {"messages": context_query, **model_config}
+        payloads = {"messages": context_query, "model": model}
 
         retry = 10
         keys = self.api_keys.copy()
@@ -783,13 +804,33 @@ class ProviderGoogleGenAI(Provider):
         self.chosen_api_key = key
         self._init_client()
 
-    async def assemble_context(self, text: str, image_urls: list[str] | None = None):
+    async def assemble_context(
+        self,
+        text: str,
+        image_urls: list[str] | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
+    ):
         """组装上下文。"""
+        # 构建内容块列表
+        content_blocks = []
+
+        # 1. 用户原始发言（OpenAI 建议：用户发言在前）
+        if text:
+            content_blocks.append({"type": "text", "text": text})
+        elif image_urls:
+            # 如果没有文本但有图片，添加占位文本
+            content_blocks.append({"type": "text", "text": "[图片]"})
+        elif extra_user_content_parts:
+            # 如果只有额外内容块，也需要添加占位文本
+            content_blocks.append({"type": "text", "text": " "})
+
+        # 2. 额外的内容块（系统提醒、指令等）
+        if extra_user_content_parts:
+            for part in extra_user_content_parts:
+                content_blocks.append(part.model_dump())
+
+        # 3. 图片内容
         if image_urls:
-            user_content = {
-                "role": "user",
-                "content": [{"type": "text", "text": text if text else "[图片]"}],
-            }
             for image_url in image_urls:
                 if image_url.startswith("http"):
                     image_path = await download_image_by_url(image_url)
@@ -802,14 +843,25 @@ class ProviderGoogleGenAI(Provider):
                 if not image_data:
                     logger.warning(f"图片 {image_url} 得到的结果为空，将忽略。")
                     continue
-                user_content["content"].append(
+                content_blocks.append(
                     {
                         "type": "image_url",
                         "image_url": {"url": image_data},
                     },
                 )
-            return user_content
+
-        return {"role": "user", "content": text}
+        # 如果只有主文本且没有额外内容块和图片，返回简单格式以保持向后兼容
+        if (
+            text
+            and not extra_user_content_parts
+            and not image_urls
+            and len(content_blocks) == 1
+            and content_blocks[0]["type"] == "text"
+        ):
+            return {"role": "user", "content": content_blocks[0]["text"]}
+
+        # 否则返回多模态格式
+        return {"role": "user", "content": content_blocks}
 
     async def encode_image_bs64(self, image_url: str) -> str:
         """将图片转换为 base64"""

astrbot/core/provider/sources/openai_source.py

@@ -17,7 +17,7 @@ from openai.types.completion_usage import CompletionUsage
 import astrbot.core.message.components as Comp
 from astrbot import logger
 from astrbot.api.provider import Provider
-from astrbot.core.agent.message import Message
+from astrbot.core.agent.message import ContentPart, Message
 from astrbot.core.agent.tool import ToolSet
 from astrbot.core.message.message_event_result import MessageChain
 from astrbot.core.provider.entities import LLMResponse, TokenUsage, ToolCallsResult
@@ -69,8 +69,7 @@ class ProviderOpenAIOfficial(Provider):
             self.client.chat.completions.create,
         ).parameters.keys()
 
-        model_config = provider_config.get("model_config", {})
+        model = provider_config.get("model", "unknown")
-        model = model_config.get("model", "unknown")
         self.set_model(model)
 
         self.reasoning_key = "reasoning_content"
@@ -349,6 +348,7 @@ class ProviderOpenAIOfficial(Provider):
         system_prompt: str | None = None,
         tool_calls_result: ToolCallsResult | list[ToolCallsResult] | None = None,
         model: str | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
         **kwargs,
     ) -> tuple:
         """准备聊天所需的有效载荷和上下文"""
@@ -356,7 +356,9 @@ class ProviderOpenAIOfficial(Provider):
             contexts = []
         new_record = None
         if prompt is not None:
-            new_record = await self.assemble_context(prompt, image_urls)
+            new_record = await self.assemble_context(
+                prompt, image_urls, extra_user_content_parts
+            )
         context_query = self._ensure_message_to_dicts(contexts)
         if new_record:
             context_query.append(new_record)
@@ -375,10 +377,9 @@ class ProviderOpenAIOfficial(Provider):
                 for tcr in tool_calls_result:
                     context_query.extend(tcr.to_openai_messages())
 
-        model_config = self.provider_config.get("model_config", {})
+        model = model or self.get_model()
-        model_config["model"] = model or self.get_model()
 
-        payloads = {"messages": context_query, **model_config}
+        payloads = {"messages": context_query, "model": model}
 
         # xAI origin search tool inject
         self._maybe_inject_xai_search(payloads, **kwargs)
@@ -478,6 +479,7 @@ class ProviderOpenAIOfficial(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ) -> LLMResponse:
         payloads, context_query = await self._prepare_chat_payload(
@@ -487,6 +489,7 @@ class ProviderOpenAIOfficial(Provider):
             system_prompt,
             tool_calls_result,
             model=model,
+            extra_user_content_parts=extra_user_content_parts,
             **kwargs,
         )
 
@@ -541,6 +544,7 @@ class ProviderOpenAIOfficial(Provider):
         system_prompt=None,
         tool_calls_result=None,
         model=None,
+        extra_user_content_parts=None,
         **kwargs,
     ) -> AsyncGenerator[LLMResponse, None]:
         """流式对话，与服务商交互并逐步返回结果"""
@@ -551,6 +555,7 @@ class ProviderOpenAIOfficial(Provider):
             system_prompt,
             tool_calls_result,
             model=model,
+            extra_user_content_parts=extra_user_content_parts,
             **kwargs,
         )
 
@@ -626,13 +631,29 @@ class ProviderOpenAIOfficial(Provider):
         self,
         text: str,
         image_urls: list[str] | None = None,
+        extra_user_content_parts: list[ContentPart] | None = None,
     ) -> dict:
         """组装成符合 OpenAI 格式的 role 为 user 的消息段"""
+        # 构建内容块列表
+        content_blocks = []
+
+        # 1. 用户原始发言（OpenAI 建议：用户发言在前）
+        if text:
+            content_blocks.append({"type": "text", "text": text})
+        elif image_urls:
+            # 如果没有文本但有图片，添加占位文本
+            content_blocks.append({"type": "text", "text": "[图片]"})
+        elif extra_user_content_parts:
+            # 如果只有额外内容块，也需要添加占位文本
+            content_blocks.append({"type": "text", "text": " "})
+
+        # 2. 额外的内容块（系统提醒、指令等）
+        if extra_user_content_parts:
+            for part in extra_user_content_parts:
+                content_blocks.append(part.model_dump())
+
+        # 3. 图片内容
         if image_urls:
-            user_content = {
-                "role": "user",
-                "content": [{"type": "text", "text": text if text else "[图片]"}],
-            }
             for image_url in image_urls:
                 if image_url.startswith("http"):
                     image_path = await download_image_by_url(image_url)
@@ -645,14 +666,25 @@ class ProviderOpenAIOfficial(Provider):
                 if not image_data:
                     logger.warning(f"图片 {image_url} 得到的结果为空，将忽略。")
                     continue
-                user_content["content"].append(
+                content_blocks.append(
                     {
                         "type": "image_url",
                         "image_url": {"url": image_data},
                     },
                 )
-            return user_content
+
-        return {"role": "user", "content": text}
+        # 如果只有主文本且没有额外内容块和图片，返回简单格式以保持向后兼容
+        if (
+            text
+            and not extra_user_content_parts
+            and not image_urls
+            and len(content_blocks) == 1
+            and content_blocks[0]["type"] == "text"
+        ):
+            return {"role": "user", "content": content_blocks[0]["text"]}
+
+        # 否则返回多模态格式
+        return {"role": "user", "content": content_blocks}
 
     async def encode_image_bs64(self, image_url: str) -> str:
         """将图片转换为 base64"""

astrbot/core/star/command_management.py

@@ -4,7 +4,7 @@ from collections import defaultdict
 from dataclasses import dataclass, field
 from typing import Any
 
-from astrbot.core import db_helper
+from astrbot.core import db_helper, logger
 from astrbot.core.db.po import CommandConfig
 from astrbot.core.star.filter.command import CommandFilter
 from astrbot.core.star.filter.command_group import CommandGroupFilter
@@ -90,6 +90,7 @@ async def toggle_command(handler_full_name: str, enabled: bool) -> CommandDescri
 async def rename_command(
     handler_full_name: str,
     new_fragment: str,
+    aliases: list[str] | None = None,
 ) -> CommandDescriptor:
     descriptor = _build_descriptor_by_full_name(handler_full_name)
     if not descriptor:
@@ -99,9 +100,24 @@ async def rename_command(
     if not new_fragment:
         raise ValueError("指令名不能为空。")
 
+    # 校验主指令名
     candidate_full = _compose_command(descriptor.parent_signature, new_fragment)
     if _is_command_in_use(handler_full_name, candidate_full):
-        raise ValueError("新的指令名已被其他指令占用，请换一个名称。")
+        raise ValueError(f"指令名 '{candidate_full}' 已被其他指令占用。")
+
+    # 校验别名
+    if aliases:
+        for alias in aliases:
+            alias = alias.strip()
+            if not alias:
+                continue
+            alias_full = _compose_command(descriptor.parent_signature, alias)
+            if _is_command_in_use(handler_full_name, alias_full):
+                raise ValueError(f"别名 '{alias_full}' 已被其他指令占用。")
+
+    existing_cfg = await db_helper.get_command_config(handler_full_name)
+    merged_extra = dict(existing_cfg.extra_data or {}) if existing_cfg else {}
+    merged_extra["resolved_aliases"] = aliases or []
 
     config = await db_helper.upsert_command_config(
         handler_full_name=handler_full_name,
@@ -114,7 +130,7 @@ async def rename_command(
         conflict_key=descriptor.original_command,
         resolution_strategy="manual_rename",
         note=None,
-        extra_data=None,
+        extra_data=merged_extra,
         auto_managed=False,
     )
     _bind_descriptor_with_config(descriptor, config)
@@ -192,12 +208,18 @@ def _collect_descriptors(include_sub_commands: bool) -> list[CommandDescriptor]:
     """收集指令，按需包含子指令。"""
     descriptors: list[CommandDescriptor] = []
     for handler in star_handlers_registry:
-        desc = _build_descriptor(handler)
+        try:
-        if not desc:
+            desc = _build_descriptor(handler)
+            if not desc:
+                continue
+            if not include_sub_commands and desc.is_sub_command:
+                continue
+            descriptors.append(desc)
+        except Exception as e:
+            logger.warning(
+                f"解析指令处理函数 {handler.handler_full_name} 失败，跳过该指令。原因: {e!s}"
+            )
             continue
-        if not include_sub_commands and desc.is_sub_command:
-            continue
-        descriptors.append(desc)
     return descriptors
 
 
@@ -357,14 +379,27 @@ def _apply_config_to_descriptor(
         new_fragment,
     )
 
+    extra = config.extra_data or {}
+    resolved_aliases = extra.get("resolved_aliases")
+    if isinstance(resolved_aliases, list):
+        descriptor.aliases = [str(x) for x in resolved_aliases if str(x).strip()]
+
 
 def _apply_config_to_runtime(
     descriptor: CommandDescriptor,
     config: CommandConfig,
 ) -> None:
     descriptor.handler.enabled = config.enabled
-    if descriptor.filter_ref and descriptor.current_fragment:
+    if descriptor.filter_ref:
-        _set_filter_fragment(descriptor.filter_ref, descriptor.current_fragment)
+        if descriptor.current_fragment:
+            _set_filter_fragment(descriptor.filter_ref, descriptor.current_fragment)
+        extra = config.extra_data or {}
+        resolved_aliases = extra.get("resolved_aliases")
+        if isinstance(resolved_aliases, list):
+            _set_filter_aliases(
+                descriptor.filter_ref,
+                [str(x) for x in resolved_aliases if str(x).strip()],
+            )
 
 
 def _bind_configs_to_descriptors(
@@ -403,6 +438,18 @@ def _set_filter_fragment(
         filter_ref._cmpl_cmd_names = None
 
 
+def _set_filter_aliases(
+    filter_ref: CommandFilter | CommandGroupFilter,
+    aliases: list[str],
+) -> None:
+    current_aliases = getattr(filter_ref, "alias", set())
+    if set(aliases) == current_aliases:
+        return
+    setattr(filter_ref, "alias", set(aliases))
+    if hasattr(filter_ref, "_cmpl_cmd_names"):
+        filter_ref._cmpl_cmd_names = None
+
+
 def _is_command_in_use(
     target_handler_full_name: str,
     candidate_full_command: str,

astrbot/core/star/context.py

@@ -267,6 +267,10 @@ class Context:
     ):
         """通过 ID 获取对应的 LLM Provider。"""
         prov = self.provider_manager.inst_map.get(provider_id)
+        if provider_id and not prov:
+            logger.warning(
+                f"没有找到 ID 为 {provider_id} 的提供商，这可能是由于您修改了提供商（模型）ID 导致的。"
+            )
         return prov
 
     def get_all_providers(self) -> list[Provider]:
@@ -296,10 +300,6 @@ class Context:
             provider_type=ProviderType.CHAT_COMPLETION,
             umo=umo,
         )
-        if prov is None:
-            raise ProviderNotFoundError(
-                "provider not found, please choose provider first"
-            )
         if not isinstance(prov, Provider):
             raise ValueError("返回的 Provider 不是 Provider 类型")
         return prov

astrbot/core/star/star_manager.py

@@ -631,7 +631,11 @@ class PluginManager:
         # 清除 pip.main 导致的多余的 logging handlers
         for handler in logging.root.handlers[:]:
             logging.root.removeHandler(handler)
-        await sync_command_configs()
+        try:
+            await sync_command_configs()
+        except Exception as e:
+            logger.error(f"同步指令配置失败: {e!s}")
+            logger.error(traceback.format_exc())
 
         if not fail_rec:
             return True, None

astrbot/core/utils/llm_metadata.py

@@ -0,0 +1,63 @@
+from typing import Literal, TypedDict
+
+import aiohttp
+
+from astrbot.core import logger
+
+
+class LLMModalities(TypedDict):
+    input: list[Literal["text", "image", "audio", "video"]]
+    output: list[Literal["text", "image", "audio", "video"]]
+
+
+class LLMLimit(TypedDict):
+    context: int
+    output: int
+
+
+class LLMMetadata(TypedDict):
+    id: str
+    reasoning: bool
+    tool_call: bool
+    knowledge: str
+    release_date: str
+    modalities: LLMModalities
+    open_weights: bool
+    limit: LLMLimit
+
+
+LLM_METADATAS: dict[str, LLMMetadata] = {}
+
+
+async def update_llm_metadata():
+    url = "https://models.dev/api.json"
+    try:
+        async with aiohttp.ClientSession() as session:
+            async with session.get(url) as response:
+                data = await response.json()
+                global LLM_METADATAS
+                models = {}
+                for info in data.values():
+                    for model in info.get("models", {}).values():
+                        model_id = model.get("id")
+                        if not model_id:
+                            continue
+                        models[model_id] = LLMMetadata(
+                            id=model_id,
+                            reasoning=model.get("reasoning", False),
+                            tool_call=model.get("tool_call", False),
+                            knowledge=model.get("knowledge", "none"),
+                            release_date=model.get("release_date", ""),
+                            modalities=model.get(
+                                "modalities", {"input": [], "output": []}
+                            ),
+                            open_weights=model.get("open_weights", False),
+                            limit=model.get("limit", {"context": 0, "output": 0}),
+                        )
+                # Replace the global cache in-place so references remain valid
+                LLM_METADATAS.clear()
+                LLM_METADATAS.update(models)
+                logger.info(f"Successfully fetched metadata for {len(models)} LLMs.")
+    except Exception as e:
+        logger.error(f"Failed to fetch LLM metadata: {e}")
+        return

astrbot/core/utils/migra_helper.py

@@ -32,6 +32,92 @@ def _migra_agent_runner_configs(conf: AstrBotConfig, ids_map: dict) -> None:
         logger.error(traceback.format_exc())
 
 
+def _migra_provider_to_source_structure(conf: AstrBotConfig) -> None:
+    """
+    Migrate old provider structure to new provider-source separation.
+    Provider only keeps: id, provider_source_id, model, modalities, custom_extra_body
+    All other fields move to provider_sources.
+    """
+    providers = conf.get("provider", [])
+    provider_sources = conf.get("provider_sources", [])
+
+    # Track if any migration happened
+    migrated = False
+
+    # Provider-only fields that should stay in provider
+    provider_only_fields = {
+        "id",
+        "provider_source_id",
+        "model",
+        "modalities",
+        "custom_extra_body",
+        "enable",
+    }
+
+    # Fields that should not go to source
+    source_exclude_fields = provider_only_fields | {"model_config"}
+
+    for provider in providers:
+        # Skip if already has provider_source_id
+        if provider.get("provider_source_id"):
+            continue
+
+        # Skip non-chat-completion types (they don't need source separation)
+        provider_type = provider.get("provider_type", "")
+        if provider_type != "chat_completion":
+            # For old types without provider_type, check type field
+            old_type = provider.get("type", "")
+            if "chat_completion" not in old_type:
+                continue
+
+        migrated = True
+        logger.info(f"Migrating provider {provider.get('id')} to new structure")
+
+        # Extract source fields from provider
+        source_fields = {}
+        for key, value in list(provider.items()):
+            if key not in source_exclude_fields:
+                source_fields[key] = value
+
+        # Create new provider_source
+        source_id = provider.get("id", "") + "_source"
+        new_source = {"id": source_id, **source_fields}
+
+        # Update provider to only keep necessary fields
+        provider["provider_source_id"] = source_id
+
+        # Extract model from model_config if exists
+        if "model_config" in provider and isinstance(provider["model_config"], dict):
+            model_config = provider["model_config"]
+            provider["model"] = model_config.get("model", "")
+
+            # Put other model_config fields into custom_extra_body
+            extra_body_fields = {k: v for k, v in model_config.items() if k != "model"}
+            if extra_body_fields:
+                if "custom_extra_body" not in provider:
+                    provider["custom_extra_body"] = {}
+                provider["custom_extra_body"].update(extra_body_fields)
+
+        # Initialize new fields if not present
+        if "modalities" not in provider:
+            provider["modalities"] = []
+        if "custom_extra_body" not in provider:
+            provider["custom_extra_body"] = {}
+
+        # Remove fields that should be in source
+        keys_to_remove = [k for k in provider.keys() if k not in provider_only_fields]
+        for key in keys_to_remove:
+            del provider[key]
+
+        # Add source to provider_sources
+        provider_sources.append(new_source)
+
+    if migrated:
+        conf["provider_sources"] = provider_sources
+        conf.save_config()
+        logger.info("Provider-source structure migration completed")
+
+
 async def migra(
     db, astrbot_config_mgr, umop_config_router, acm: AstrBotConfigManager
 ) -> None:
@@ -71,3 +157,10 @@ async def migra(
 
     for conf in acm.confs.values():
         _migra_agent_runner_configs(conf, ids_map)
+
+    # Migrate providers to new structure: extract source fields to provider_sources
+    try:
+        _migra_provider_to_source_structure(astrbot_config)
+    except Exception as e:
+        logger.error(f"Migration for provider-source structure failed: {e!s}")
+        logger.error(traceback.format_exc())

astrbot/dashboard/routes/chat.py

@@ -436,7 +436,7 @@ class ChatRoute(Route):
                             accumulated_parts = []
                             accumulated_text = ""
                             accumulated_reasoning = ""
-                            tool_calls = {}
+                            # tool_calls = {}
                             agent_stats = {}
             except BaseException as e:
                 logger.exception(f"WebChat stream unexpected error: {e}", exc_info=True)

astrbot/dashboard/routes/command.py

@@ -61,12 +61,13 @@ class CommandRoute(Route):
         data = await request.get_json()
         handler_full_name = data.get("handler_full_name")
         new_name = data.get("new_name")
+        aliases = data.get("aliases")
 
         if not handler_full_name or not new_name:
             return Response().error("handler_full_name 与 new_name 均为必填。").__dict__
 
         try:
-            await rename_command_service(handler_full_name, new_name)
+            await rename_command_service(handler_full_name, new_name, aliases=aliases)
         except ValueError as exc:
             return Response().error(str(exc)).__dict__
 

astrbot/dashboard/routes/config.py

@@ -6,7 +6,7 @@ from typing import Any
 
 from quart import request
 
-from astrbot.core import file_token_service, logger
+from astrbot.core import astrbot_config, file_token_service, logger
 from astrbot.core.config.astrbot_config import AstrBotConfig
 from astrbot.core.config.default import (
     CONFIG_METADATA_2,
@@ -21,6 +21,7 @@ from astrbot.core.platform.register import platform_cls_map, platform_registry
 from astrbot.core.provider import Provider
 from astrbot.core.provider.register import provider_registry
 from astrbot.core.star.star import star_registry
+from astrbot.core.utils.llm_metadata import LLM_METADATAS
 from astrbot.core.utils.webhook_utils import ensure_platform_webhook_config
 
 from .route import Response, Route, RouteContext
@@ -179,13 +180,157 @@ class ConfigRoute(Route):
             "/config/provider/new": ("POST", self.post_new_provider),
             "/config/provider/update": ("POST", self.post_update_provider),
             "/config/provider/delete": ("POST", self.post_delete_provider),
+            "/config/provider/template": ("GET", self.get_provider_template),
             "/config/provider/check_one": ("GET", self.check_one_provider_status),
             "/config/provider/list": ("GET", self.get_provider_config_list),
             "/config/provider/model_list": ("GET", self.get_provider_model_list),
             "/config/provider/get_embedding_dim": ("POST", self.get_embedding_dim),
+            "/config/provider_sources/models": (
+                "GET",
+                self.get_provider_source_models,
+            ),
+            "/config/provider_sources/update": (
+                "POST",
+                self.update_provider_source,
+            ),
+            "/config/provider_sources/delete": (
+                "POST",
+                self.delete_provider_source,
+            ),
         }
         self.register_routes()
 
+    async def delete_provider_source(self):
+        """删除 provider_source，并更新关联的 providers"""
+        post_data = await request.json
+        if not post_data:
+            return Response().error("缺少配置数据").__dict__
+
+        provider_source_id = post_data.get("id")
+        if not provider_source_id:
+            return Response().error("缺少 provider_source_id").__dict__
+
+        provider_sources = self.config.get("provider_sources", [])
+        target_idx = next(
+            (
+                i
+                for i, ps in enumerate(provider_sources)
+                if ps.get("id") == provider_source_id
+            ),
+            -1,
+        )
+
+        if target_idx == -1:
+            return Response().error("未找到对应的 provider source").__dict__
+
+        # 删除 provider_source
+        del provider_sources[target_idx]
+
+        # 写回配置
+        self.config["provider_sources"] = provider_sources
+
+        # 删除引用了该 provider_source 的 providers
+        await self.core_lifecycle.provider_manager.delete_provider(
+            provider_source_id=provider_source_id
+        )
+
+        try:
+            save_config(self.config, self.config, is_core=True)
+        except Exception as e:
+            logger.error(traceback.format_exc())
+            return Response().error(str(e)).__dict__
+
+        return Response().ok(message="删除 provider source 成功").__dict__
+
+    async def update_provider_source(self):
+        """更新或新增 provider_source，并重载关联的 providers"""
+        post_data = await request.json
+        if not post_data:
+            return Response().error("缺少配置数据").__dict__
+
+        new_source_config = post_data.get("config") or post_data
+        original_id = post_data.get("original_id")
+        if not original_id:
+            return Response().error("缺少 original_id").__dict__
+
+        if not isinstance(new_source_config, dict):
+            return Response().error("缺少或错误的配置数据").__dict__
+
+        # 确保配置中有 id 字段
+        if not new_source_config.get("id"):
+            new_source_config["id"] = original_id
+
+        provider_sources = self.config.get("provider_sources", [])
+
+        for ps in provider_sources:
+            if ps.get("id") == new_source_config["id"] and ps.get("id") != original_id:
+                return (
+                    Response()
+                    .error(
+                        f"Provider source ID '{new_source_config['id']}' exists already, please try another ID.",
+                    )
+                    .__dict__
+                )
+
+        # 查找旧的 provider_source，若不存在则追加为新配置
+        target_idx = next(
+            (i for i, ps in enumerate(provider_sources) if ps.get("id") == original_id),
+            -1,
+        )
+
+        old_id = original_id
+        if target_idx == -1:
+            provider_sources.append(new_source_config)
+        else:
+            old_id = provider_sources[target_idx].get("id")
+            provider_sources[target_idx] = new_source_config
+
+        # 更新引用了该 provider_source 的 providers
+        affected_providers = []
+        for provider in self.config.get("provider", []):
+            if provider.get("provider_source_id") == old_id:
+                provider["provider_source_id"] = new_source_config["id"]
+                affected_providers.append(provider)
+
+        # 写回配置
+        self.config["provider_sources"] = provider_sources
+
+        try:
+            save_config(self.config, self.config, is_core=True)
+        except Exception as e:
+            logger.error(traceback.format_exc())
+            return Response().error(str(e)).__dict__
+
+        # 重载受影响的 providers，使新的 source 配置生效
+        reload_errors = []
+        prov_mgr = self.core_lifecycle.provider_manager
+        for provider in affected_providers:
+            try:
+                await prov_mgr.reload(provider)
+            except Exception as e:
+                logger.error(traceback.format_exc())
+                reload_errors.append(f"{provider.get('id')}: {e}")
+
+        if reload_errors:
+            return (
+                Response()
+                .error("更新成功，但部分提供商重载失败: " + ", ".join(reload_errors))
+                .__dict__
+            )
+
+        return Response().ok(message="更新 provider source 成功").__dict__
+
+    async def get_provider_template(self):
+        config_schema = {
+            "provider": CONFIG_METADATA_2["provider_group"]["metadata"]["provider"]
+        }
+        data = {
+            "config_schema": config_schema,
+            "providers": astrbot_config["provider"],
+            "provider_sources": astrbot_config["provider_sources"],
+        }
+        return Response().ok(data=data).__dict__
+
     async def get_uc_table(self):
         """获取 UMOP 配置路由表"""
         return Response().ok({"routing": self.ucr.umop_to_conf_id}).__dict__
@@ -433,9 +578,25 @@ class ConfigRoute(Route):
             return Response().error("缺少参数 provider_type").__dict__
         provider_type_ls = provider_type.split(",")
         provider_list = []
-        astrbot_config = self.core_lifecycle.astrbot_config
+        ps = self.core_lifecycle.provider_manager.providers_config
-        for provider in astrbot_config["provider"]:
+        p_source_pt = {
-            if provider.get("provider_type", None) in provider_type_ls:
+            psrc["id"]: psrc["provider_type"]
+            for psrc in self.core_lifecycle.provider_manager.provider_sources_config
+        }
+        for provider in ps:
+            ps_id = provider.get("provider_source_id", None)
+            if (
+                ps_id
+                and ps_id in p_source_pt
+                and p_source_pt[ps_id] in provider_type_ls
+            ):
+                # chat
+                prov = self.core_lifecycle.provider_manager.get_merged_provider_config(
+                    provider
+                )
+                provider_list.append(prov)
+            elif not ps_id and provider.get("provider_type", None) in provider_type_ls:
+                # agent runner, embedding, etc
                 provider_list.append(provider)
         return Response().ok(provider_list).__dict__
 
@@ -458,9 +619,18 @@ class ConfigRoute(Route):
 
         try:
             models = await provider.get_models()
+            models = models or []
+
+            metadata_map = {}
+            for model_id in models:
+                meta = LLM_METADATAS.get(model_id)
+                if meta:
+                    metadata_map[model_id] = meta
+
             ret = {
                 "models": models,
                 "provider_id": provider_id,
+                "model_metadata": metadata_map,
             }
             return Response().ok(ret).__dict__
         except Exception as e:
@@ -522,6 +692,104 @@ class ConfigRoute(Route):
             logger.error(traceback.format_exc())
             return Response().error(f"获取嵌入维度失败: {e!s}").__dict__
 
+    async def get_provider_source_models(self):
+        """获取指定 provider_source 支持的模型列表
+
+        本质上会临时初始化一个 Provider 实例，调用 get_models() 获取模型列表，然后销毁实例
+        """
+        provider_source_id = request.args.get("source_id")
+        if not provider_source_id:
+            return Response().error("缺少参数 source_id").__dict__
+
+        try:
+            from astrbot.core.provider.register import provider_cls_map
+
+            # 从配置中查找对应的 provider_source
+            provider_sources = self.config.get("provider_sources", [])
+            provider_source = None
+            for ps in provider_sources:
+                if ps.get("id") == provider_source_id:
+                    provider_source = ps
+                    break
+
+            if not provider_source:
+                return (
+                    Response()
+                    .error(f"未找到 ID 为 {provider_source_id} 的 provider_source")
+                    .__dict__
+                )
+
+            # 获取 provider 类型
+            provider_type = provider_source.get("type", None)
+            if not provider_type:
+                return Response().error("provider_source 缺少 type 字段").__dict__
+
+            try:
+                self.core_lifecycle.provider_manager.dynamic_import_provider(
+                    provider_type
+                )
+            except ImportError as e:
+                logger.error(traceback.format_exc())
+                return Response().error(f"动态导入提供商适配器失败: {e!s}").__dict__
+
+            # 获取对应的 provider 类
+            if provider_type not in provider_cls_map:
+                return (
+                    Response()
+                    .error(f"未找到适用于 {provider_type} 的提供商适配器")
+                    .__dict__
+                )
+
+            provider_metadata = provider_cls_map[provider_type]
+            cls_type = provider_metadata.cls_type
+
+            if not cls_type:
+                return Response().error(f"无法找到 {provider_type} 的类").__dict__
+
+            # 检查是否是 Provider 类型
+            if not issubclass(cls_type, Provider):
+                return (
+                    Response()
+                    .error(f"提供商 {provider_type} 不支持获取模型列表")
+                    .__dict__
+                )
+
+            # 临时实例化 provider
+            inst = cls_type(provider_source, {})
+
+            # 如果有 initialize 方法，调用它
+            init_fn = getattr(inst, "initialize", None)
+            if inspect.iscoroutinefunction(init_fn):
+                await init_fn()
+
+            # 获取模型列表
+            models = await inst.get_models()
+            models = models or []
+
+            metadata_map = {}
+            for model_id in models:
+                meta = LLM_METADATAS.get(model_id)
+                if meta:
+                    metadata_map[model_id] = meta
+
+            # 销毁实例（如果有 terminate 方法）
+            terminate_fn = getattr(inst, "terminate", None)
+            if inspect.iscoroutinefunction(terminate_fn):
+                await terminate_fn()
+
+            logger.info(
+                f"获取到 provider_source {provider_source_id} 的模型列表: {models}",
+            )
+
+            return (
+                Response()
+                .ok({"models": models, "model_metadata": metadata_map})
+                .__dict__
+            )
+        except Exception as e:
+            logger.error(traceback.format_exc())
+            return Response().error(f"获取模型列表失败: {e!s}").__dict__
+
     async def get_platform_list(self):
         """获取所有平台的列表"""
         platform_list = []
@@ -533,7 +801,15 @@ class ConfigRoute(Route):
         data = await request.json
         config = data.get("config", None)
         conf_id = data.get("conf_id", None)
+
         try:
+            # 不更新 provider_sources, provider, platform
+            # 这些配置有单独的接口进行更新
+            if conf_id == "default":
+                no_update_keys = ["provider_sources", "provider", "platform"]
+                for key in no_update_keys:
+                    config[key] = self.acm.default_conf[key]
+
             await self._save_astrbot_configs(config, conf_id)
             await self.core_lifecycle.reload_pipeline_scheduler(conf_id)
             return Response().ok(None, "保存成功~").__dict__
@@ -573,28 +849,30 @@ class ConfigRoute(Route):
 
     async def post_new_provider(self):
         new_provider_config = await request.json
-        self.config["provider"].append(new_provider_config)
+
         try:
-            save_config(self.config, self.config, is_core=True)
+            await self.core_lifecycle.provider_manager.create_provider(
-            await self.core_lifecycle.provider_manager.load_provider(
+                new_provider_config
-                new_provider_config,
             )
         except Exception as e:
             return Response().error(str(e)).__dict__
-        return Response().ok(None, "新增服务提供商配置成功~").__dict__
+        return Response().ok(None, "新增服务提供商配置成功").__dict__
 
     async def post_update_platform(self):
         update_platform_config = await request.json
-        platform_id = update_platform_config.get("id", None)
+        origin_platform_id = update_platform_config.get("id", None)
         new_config = update_platform_config.get("config", None)
-        if not platform_id or not new_config:
+        if not origin_platform_id or not new_config:
             return Response().error("参数错误").__dict__
 
+        if origin_platform_id != new_config.get("id", None):
+            return Response().error("机器人名称不允许修改").__dict__
+
         # 如果是支持统一 webhook 模式的平台，且启用了统一 webhook 模式，确保有 webhook_uuid
         ensure_platform_webhook_config(new_config)
 
         for i, platform in enumerate(self.config["platform"]):
-            if platform["id"] == platform_id:
+            if platform["id"] == origin_platform_id:
                 self.config["platform"][i] = new_config
                 break
         else:
@@ -609,21 +887,15 @@ class ConfigRoute(Route):
 
     async def post_update_provider(self):
         update_provider_config = await request.json
-        provider_id = update_provider_config.get("id", None)
+        origin_provider_id = update_provider_config.get("id", None)
         new_config = update_provider_config.get("config", None)
-        if not provider_id or not new_config:
+        if not origin_provider_id or not new_config:
             return Response().error("参数错误").__dict__
 
-        for i, provider in enumerate(self.config["provider"]):
-            if provider["id"] == provider_id:
-                self.config["provider"][i] = new_config
-                break
-        else:
-            return Response().error("未找到对应服务提供商").__dict__
-
         try:
-            save_config(self.config, self.config, is_core=True)
+            await self.core_lifecycle.provider_manager.update_provider(
-            await self.core_lifecycle.provider_manager.reload(new_config)
+                origin_provider_id, new_config
+            )
         except Exception as e:
             return Response().error(str(e)).__dict__
         return Response().ok(None, "更新成功，已经实时生效~").__dict__
@@ -646,19 +918,17 @@ class ConfigRoute(Route):
 
     async def post_delete_provider(self):
         provider_id = await request.json
-        provider_id = provider_id.get("id")
+        provider_id = provider_id.get("id", "")
-        for i, provider in enumerate(self.config["provider"]):
+        if not provider_id:
-            if provider["id"] == provider_id:
+            return Response().error("缺少参数 id").__dict__
-                del self.config["provider"][i]
+
-                break
-        else:
-            return Response().error("未找到对应服务提供商").__dict__
         try:
-            save_config(self.config, self.config, is_core=True)
+            await self.core_lifecycle.provider_manager.delete_provider(
-            await self.core_lifecycle.provider_manager.terminate_provider(provider_id)
+                provider_id=provider_id
+            )
         except Exception as e:
             return Response().error(str(e)).__dict__
-        return Response().ok(None, "删除成功，已经实时生效~").__dict__
+        return Response().ok(None, "删除成功，已经实时生效。").__dict__
 
     async def get_llm_tools(self):
         """获取函数调用工具。包含了本地加载的以及 MCP 服务的工具"""

astrbot/dashboard/routes/stat.py

@@ -1,6 +1,9 @@
+import os
+import re
 import threading
 import time
 import traceback
+from functools import cmp_to_key
 
 import aiohttp
 import psutil
@@ -11,7 +14,9 @@ from astrbot.core.config import VERSION
 from astrbot.core.core_lifecycle import AstrBotCoreLifecycle
 from astrbot.core.db import BaseDatabase
 from astrbot.core.db.migration.helper import check_migration_needed_v4
+from astrbot.core.utils.astrbot_path import get_astrbot_path
 from astrbot.core.utils.io import get_dashboard_version
+from astrbot.core.utils.version_comparator import VersionComparator
 
 from .route import Response, Route, RouteContext
 
@@ -30,6 +35,8 @@ class StatRoute(Route):
             "/stat/start-time": ("GET", self.get_start_time),
             "/stat/restart-core": ("POST", self.restart_core),
             "/stat/test-ghproxy-connection": ("POST", self.test_ghproxy_connection),
+            "/stat/changelog": ("GET", self.get_changelog),
+            "/stat/changelog/list": ("GET", self.list_changelog_versions),
         }
         self.db_helper = db_helper
         self.register_routes()
@@ -183,3 +190,92 @@ class StatRoute(Route):
         except Exception as e:
             logger.error(traceback.format_exc())
             return Response().error(f"Error: {e!s}").__dict__
+
+    async def get_changelog(self):
+        """获取指定版本的更新日志"""
+        try:
+            version = request.args.get("version")
+            if not version:
+                return Response().error("version parameter is required").__dict__
+
+            version = version.lstrip("v")
+
+            # 防止路径遍历攻击
+            if not re.match(r"^[a-zA-Z0-9._-]+$", version):
+                return Response().error("Invalid version format").__dict__
+            if ".." in version or "/" in version or "\\" in version:
+                return Response().error("Invalid version format").__dict__
+
+            filename = f"v{version}.md"
+            project_path = get_astrbot_path()
+            changelogs_dir = os.path.join(project_path, "changelogs")
+            changelog_path = os.path.join(changelogs_dir, filename)
+
+            # 规范化路径，防止符号链接攻击
+            changelog_path = os.path.realpath(changelog_path)
+            changelogs_dir = os.path.realpath(changelogs_dir)
+
+            # 验证最终路径在预期的 changelogs 目录内（防止路径遍历）
+            # 确保规范化后的路径以 changelogs_dir 开头，且是目录内的文件
+            changelog_path_normalized = os.path.normpath(changelog_path)
+            changelogs_dir_normalized = os.path.normpath(changelogs_dir)
+
+            # 检查路径是否在预期目录内（必须是目录的子文件，不能是目录本身）
+            expected_prefix = changelogs_dir_normalized + os.sep
+            if not changelog_path_normalized.startswith(expected_prefix):
+                logger.warning(
+                    f"Path traversal attempt detected: {version} -> {changelog_path}",
+                )
+                return Response().error("Invalid version format").__dict__
+
+            if not os.path.exists(changelog_path):
+                return (
+                    Response()
+                    .error(f"Changelog for version {version} not found")
+                    .__dict__
+                )
+            if not os.path.isfile(changelog_path):
+                return (
+                    Response()
+                    .error(f"Changelog for version {version} not found")
+                    .__dict__
+                )
+
+            with open(changelog_path, encoding="utf-8") as f:
+                content = f.read()
+
+            return Response().ok({"content": content, "version": version}).__dict__
+        except Exception as e:
+            logger.error(traceback.format_exc())
+            return Response().error(f"Error: {e!s}").__dict__
+
+    async def list_changelog_versions(self):
+        """获取所有可用的更新日志版本列表"""
+        try:
+            project_path = get_astrbot_path()
+            changelogs_dir = os.path.join(project_path, "changelogs")
+
+            if not os.path.exists(changelogs_dir):
+                return Response().ok({"versions": []}).__dict__
+
+            versions = []
+            for filename in os.listdir(changelogs_dir):
+                if filename.endswith(".md") and filename.startswith("v"):
+                    # 提取版本号（去除 v 前缀和 .md 后缀）
+                    version = filename[1:-3]  # 去掉 "v" 和 ".md"
+                    # 验证版本号格式
+                    if re.match(r"^[a-zA-Z0-9._-]+$", version):
+                        versions.append(version)
+
+            # 按版本号排序（降序，最新的在前）
+            # 使用项目中的 VersionComparator 进行语义化版本号排序
+            versions.sort(
+                key=cmp_to_key(
+                    lambda v1, v2: VersionComparator.compare_version(v2, v1),
+                ),
+            )
+
+            return Response().ok({"versions": versions}).__dict__
+        except Exception as e:
+            logger.error(traceback.format_exc())
+            return Response().error(f"Error: {e!s}").__dict__

build_nuitka.py

@@ -1,134 +0,0 @@
-#!/usr/bin/env python3
-"""
-Use Nuitka to build the AstrBot project into standalone executables
-"""
-
-import os
-import platform
-import subprocess
-import sys
-from pathlib import Path
-
-
-def get_platform_info():
-    """fetch the current platform information"""
-    system = platform.system()
-    machine = platform.machine()
-    return system, machine
-
-
-def build_with_nuitka():
-    """use Nuitka to build the project"""
-    system, machine = get_platform_info()
-
-    print(f"🚀 Starting build for {system} ({machine}) platform...")
-
-    # Output directory
-    output_dir = Path("build/nuitka")
-    output_dir.mkdir(parents=True, exist_ok=True)
-
-    # Base Nuitka command
-    nuitka_cmd = [
-        sys.executable,
-        "-m",
-        "nuitka",
-        "--standalone",  # Create standalone directory
-        "--onefile",  # Single file mode
-        "--follow-imports",  # Follow all imports
-        "--enable-plugin=multiprocessing",  # Enable multiprocessing support
-        "--output-dir=build/nuitka",  # Output directory
-        "--quiet",  # Reduce output verbosity
-        "--assume-yes-for-downloads",  # Automatically download dependencies
-        "--jobs=4",  # Use multiple CPU cores
-    ]
-
-    # include specific packages
-    include_packages = [
-        "astrbot",
-    ]
-
-    for pkg in include_packages:
-        nuitka_cmd.extend([f"--include-package={pkg}"])
-
-    # include data directories
-    # data_includes = [
-    #     "data/config",
-    #     "data/plugins",
-    #     "data/temp",
-    # ]
-
-    # for data_dir in data_includes:
-    #     if os.path.exists(data_dir):
-    #         nuitka_cmd.extend([f"--include-data-dir={data_dir}={data_dir}"])
-
-    # include packages directory (built-in plugins)
-    # if os.path.exists("packages"):
-    #     nuitka_cmd.extend(["--include-data-dir=packages=packages"])
-
-    # Platform specific settings
-    if system == "Darwin":  # macOS
-        nuitka_cmd.extend(
-            [
-                "--macos-create-app-bundle",  # Create .app bundle
-                "--macos-app-name=AstrBot",
-            ]
-        )
-        # macOS icon (if exists)
-        icon_path = "dashboard/src-tauri/icons/icon.icns"
-        if os.path.exists(icon_path):
-            nuitka_cmd.extend([f"--macos-app-icon={icon_path}"])
-    elif system == "Windows":
-        nuitka_cmd.extend(
-            [
-                "--windows-console-mode=disable",  # 无控制台窗口
-            ]
-        )
-        # Windows icon (if exists)
-        icon_path = "dashboard/src-tauri/icons/icon.ico"
-        if os.path.exists(icon_path):
-            nuitka_cmd.extend([f"--windows-icon-from-ico={icon_path}"])
-
-    # Main file to compile
-    nuitka_cmd.append("main.py")
-
-    print(f"📦 Executing command: {' '.join(nuitka_cmd)}")
-
-    try:
-        subprocess.run(nuitka_cmd, check=True)
-        print("✅ Nuitka build successful!")
-
-        # Find the generated executable
-        if system == "Darwin":
-            built_file = list(output_dir.glob("*.app"))
-            if built_file:
-                print(f"Generated macOS app: {built_file[0]}")
-        elif system == "Windows":
-            built_file = list(output_dir.glob("*.exe"))
-            if built_file:
-                print(f"Generated Windows executable: {built_file[0]}")
-        else:  # Linux
-            built_file = list(output_dir.glob("main.bin"))
-            if built_file:
-                print(f"Generated Linux executable: {built_file[0]}")
-
-        return True
-    except subprocess.CalledProcessError as e:
-        print(f"❌ Nuitka build failed: {e}")
-        return False
-
-
-if __name__ == "__main__":
-    print("=" * 60)
-    print("AstrBot Nuitka Builder")
-    print("=" * 60)
-
-    # 构建
-    if build_with_nuitka():
-        print("\n" + "=" * 60)
-        print("🎉 Build Complete!")
-        print("=" * 60)
-    else:
-        print("\n" + "=" * 60)
-        print("❌ Build Failed")
-        print("=" * 60)
-        sys.exit(1)

build_pyinstaller.py

@@ -1,134 +0,0 @@
-#!/usr/bin/env python3
-"""
-Use PyInstaller to build the AstrBot project into standalone executables
-"""
-
-import platform
-import subprocess
-import sys
-from pathlib import Path
-
-
-def get_platform_info():
-    """fetch the current platform information"""
-    system = platform.system()
-    machine = platform.machine()
-    return system, machine
-
-
-def build_with_pyinstaller():
-    """use PyInstaller to build the project"""
-    system, machine = get_platform_info()
-
-    print(f"🚀 Starting build for {system} ({machine}) platform...")
-
-    # Output directory
-    output_dir = Path("build/pyinstaller")
-    output_dir.mkdir(parents=True, exist_ok=True)
-
-    # Base PyInstaller command
-    pyinstaller_cmd = [
-        sys.executable,
-        "-m",
-        "PyInstaller",
-        "--clean",  # Clean cache before build
-        "--noconfirm",  # Replace output directory without asking
-        "--onefile",  # Single file mode
-        "--distpath=build/pyinstaller/dist",  # Distribution directory
-        "--workpath=build/pyinstaller/build",  # Work directory
-        "--specpath=build/pyinstaller",  # Spec file directory
-        "--name=AstrBot",  # Output executable name
-    ]
-    # Platform specific settings
-    # if system == "Darwin":  # macOS
-    #     # macOS icon (if exists)
-    #     icon_path = "dashboard/src-tauri/icons/icon.icns"
-    #     if os.path.exists(icon_path):
-    #         pyinstaller_cmd.extend([f"--icon={icon_path}"])
-    #     # Create .app bundle
-    #     pyinstaller_cmd.extend(["--windowed"])
-    # elif system == "Windows":
-    #     # Windows icon (if exists)
-    #     icon_path = "dashboard/src-tauri/icons/icon.ico"
-    #     if os.path.exists(icon_path):
-    #         pyinstaller_cmd.extend([f"--icon={icon_path}"])
-    #     # No console window
-    #     pyinstaller_cmd.extend(["--windowed"])
-    # else:  # Linux
-    #     pyinstaller_cmd.extend(["--console"])
-
-    # Main file to compile
-    pyinstaller_cmd.append("main.py")
-
-    print(f"📦 Executing command: {' '.join(pyinstaller_cmd)}")
-
-    try:
-        subprocess.run(pyinstaller_cmd, check=True)
-        print("✅ PyInstaller build successful!")
-
-        # Find the generated executable
-        dist_dir = output_dir / "dist"
-        if system == "Darwin":
-            built_file = list(dist_dir.glob("AstrBot.app"))
-            if not built_file:
-                built_file = list(dist_dir.glob("AstrBot"))
-            if built_file:
-                print(f"📱 Generated macOS app: {built_file[0]}")
-        elif system == "Windows":
-            built_file = list(dist_dir.glob("AstrBot.exe"))
-            if built_file:
-                print(f"💻 Generated Windows executable: {built_file[0]}")
-        else:  # Linux
-            built_file = list(dist_dir.glob("AstrBot"))
-            if built_file:
-                print(f"🐧 Generated Linux executable: {built_file[0]}")
-
-        print(f"\n📁 Output directory: {dist_dir.absolute()}")
-        return True
-    except subprocess.CalledProcessError as e:
-        print(f"❌ PyInstaller build failed: {e}")
-        return False
-    except Exception as e:
-        print(f"❌ Unexpected error: {e}")
-        return False
-
-
-def install_pyinstaller():
-    """Install PyInstaller if not already installed"""
-    try:
-        import PyInstaller
-
-        print(f"✅ PyInstaller already installed (version {PyInstaller.__version__})")
-        return True
-    except ImportError:
-        print("📥 PyInstaller not found, installing...")
-        try:
-            subprocess.run(
-                [sys.executable, "-m", "pip", "install", "pyinstaller"], check=True
-            )
-            print("✅ PyInstaller installed successfully!")
-            return True
-        except subprocess.CalledProcessError as e:
-            print(f"❌ Failed to install PyInstaller: {e}")
-            return False
-
-
-if __name__ == "__main__":
-    print("=" * 60)
-    print("AstrBot PyInstaller Builder")
-    print("=" * 60)
-
-    # Check and install PyInstaller
-    if not install_pyinstaller():
-        sys.exit(1)
-
-    # Build
-    if build_with_pyinstaller():
-        print("\n" + "=" * 60)
-        print("🎉 Build Complete!")
-        print("=" * 60)
-    else:
-        print("\n" + "=" * 60)
-        print("❌ Build Failed")
-        print("=" * 60)
-        sys.exit(1)

changelogs/v4.10.0-alpha.1.md

@@ -0,0 +1,34 @@
+## What's Changed
+
+> 📢 在升级前，请**完整阅读**本次更新日志。
+> 
+> **特别提醒：**
+> 1. 该版本为 alpha.1 预览版本。
+> 2. 本次升级**如果再降级**，会由于提供商配置的变更，导致提供商配置错乱，需要手动删除后重新添加。
+> 3. 此版本 WebUI 包体相较上一个版本增加约 **193%**，共约 **9.8 MB**，升级可能会需要一些时间。
+
+### 重构与优化
+
+- 重构 Provider 页面和提供商的配置结构，将 Chat Provider 配置拆分为 Provider Source（提供商源）和 Provider（代表提供商源的各个模型），引入了提供商模型自动发现、模型元数据自动发现的功能，**提供更加便捷的模型添加体验**。
+- ⚠️ 将 “MCP” 页面移动到了 “插件” 页面中
+- ⚠️ 将 “MCP” 页面中的工具管理移动到了 “插件” -> “管理行为” 中。
+- ⚠️ 将 “QQ 个人号（OneBot v11）” 机器人适配器类型更名为 “OneBot v11”，并将其 Logo 更改为 OneBot 的 Logo。
+- ⚠️ AstrBot WebChat 升级为 **AstrBot ChatUI**，入口从边栏修改为顶部（右上角）切换按钮。
+- 优化引用消息的逻辑，减少对模型输入缓存的破坏。
+
+### 修复
+
+- ‼️ 修复部分情况下，分段回复无法正常分段的问题。
+- 修复处理工具返回结果的过程中，导致一些直接发送图片的工具（如生图工具）无法正确发送到用户的问题。
+- 修复 WebChat 部分情况下，上一条消息文字内容增量到下一条消息的问题。
+
+### 新增
+
+- 支持**指令管理**，设置指令别名、解决指令冲突、查看指令详情等。入口：“插件” -> “管理行为”。
+- 支持 Google Gemini 3 系列引入的 [Thinking Level](https://ai.google.dev/gemini-api/docs/thinking#thinking-levels) 配置。
+- 支持记录每条 LLM 消息的耗时、Token 使用量、TTFT 数据，以及每次 Agent Loop 的各种统计数据。
+- AstrBot ChatUI 支持查看每条消息的 TTFT、Token 使用量数据。
+- AstrBot ChatUI 支持显示每次工具调用的耗时、参数和响应。
+- AstrBot ChatUI 支持渲染 Mermaid、LateX 内容，优化了 Code Block 的显示效果（使用 Monaco Editor），并减少 DOM 更新于内存占用。（Powered by [Simon-He95/markstream-vue](https://github.com/Simon-He95/markstream-vue)）
+- 支持查看 Changelog 历史版本更新日志。
+- 🎄

changelogs/v4.10.0-alpha.2.md

@@ -0,0 +1,44 @@
+## What's Changed
+
+> 📢 在升级前，请**完整阅读**本次更新日志。
+> 
+> **特别提醒：**
+> 1. 该版本为 alpha.2 预览版本。
+> 2. 本次升级**如果再降级**，会由于提供商配置的变更，导致提供商配置错乱，需要手动删除后重新添加。
+> 3. 此版本 WebUI 包体相较上一个版本增加约 **193%**，共约 **9.8 MB**，升级可能会需要一些时间。
+
+## alpha.1 -> alpha.2
+
+- 修复：“对话数据”页对话轨迹详情显示异常的问题
+- 优化：当 Agent 达到最大步数时的处理。在达到最大步数后，会移除所有请求中的 tools 并告知模型根据上下文进行最终总结。
+- 优化：LLM tools 执行的错误处理，减少工具调用无限循环的问题。
+- 优化：ChatUI 打开模型选择菜单时，会重新获取提供商配置。
+- 优化：ChatUI 新建对话并发送消息后，对话列表页自动选中该对话。
+
+## 4.10.0 变化
+
+### 重构与优化
+
+- 重构 Provider 页面和提供商的配置结构，将 Chat Provider 配置拆分为 Provider Source（提供商源）和 Provider（代表提供商源的各个模型），引入了提供商模型自动发现、模型元数据自动发现的功能，**提供更加便捷的模型添加体验**。
+- ⚠️ 将 “MCP” 页面移动到了 “插件” 页面中
+- ⚠️ 将 “MCP” 页面中的工具管理移动到了 “插件” -> “管理行为” 中。
+- ⚠️ 将 “QQ 个人号（OneBot v11）” 机器人适配器类型更名为 “OneBot v11”，并将其 Logo 更改为 OneBot 的 Logo。
+- ⚠️ AstrBot WebChat 升级为 **AstrBot ChatUI**，入口从边栏修改为顶部（右上角）切换按钮。
+- 优化引用消息的逻辑，减少对模型输入缓存的破坏。
+
+### 修复
+
+- ‼️ 修复部分情况下，分段回复无法正常分段的问题。
+- 修复处理工具返回结果的过程中，导致一些直接发送图片的工具（如生图工具）无法正确发送到用户的问题。
+- 修复 WebChat 部分情况下，上一条消息文字内容增量到下一条消息的问题。
+
+### 新增
+
+- 支持**指令管理**，设置指令别名、解决指令冲突、查看指令详情等。入口：“插件” -> “管理行为”。
+- 支持 Google Gemini 3 系列引入的 [Thinking Level](https://ai.google.dev/gemini-api/docs/thinking#thinking-levels) 配置。
+- 支持记录每条 LLM 消息的耗时、Token 使用量、TTFT 数据，以及每次 Agent Loop 的各种统计数据。
+- AstrBot ChatUI 支持查看每条消息的 TTFT、Token 使用量数据。
+- AstrBot ChatUI 支持显示每次工具调用的耗时、参数和响应。
+- AstrBot ChatUI 支持渲染 Mermaid、LateX 内容，优化了 Code Block 的显示效果（使用 Monaco Editor），并减少 DOM 更新于内存占用。（Powered by [Simon-He95/markstream-vue](https://github.com/Simon-He95/markstream-vue)）
+- 支持查看 Changelog 历史版本更新日志。
+- 🎄

changelogs/v4.10.0.md

@@ -0,0 +1,40 @@
+## What's Changed
+
+> 📢 在升级前，请**完整阅读**本次更新日志。
+> 
+> **特别提醒：**
+> 1. 本次升级**如果再降级**，会由于提供商配置的变更，导致提供商配置错乱，需要手动删除后重新添加。
+> 2. 此版本 WebUI 包体相较上一个版本增加约 **193%**，共约 **9.8 MB**，升级可能会需要一些时间。
+> 3. **升级后请务必确保 WebUI 和 AstrBot Core 版本一致**，否则会产生预期之外的情况。（判断方法：日志中出现 `WebUI 版本已是最新。` 即为一致的版本，`检测到 WebUI 版本 (xxx) 与当前 AstrBot 版本 (xxx) 不符。` 即为不一致的版本。此版本的判断方法也可通查看 WebUI 右上角是否出现 Bot / Chat 的切换按钮控件来判断是否是新版本的 WebUI）。
+> 4. 如果有任何问题请提交 [Issue](https://github.com/AstrBotDevs/AstrBot/issues) 并附带 `v4.10.0` tag。
+
+### 重构与优化
+
+- 重构 Provider 页面和提供商的配置结构，将 Chat Provider 配置拆分为 Provider Source（提供商源）和 Provider（代表提供商源的各个模型），引入了提供商模型自动发现、模型元数据自动发现的功能，**提供更加便捷的模型添加体验**。
+- ⚠️ 将 “MCP” 页面移动到了 “插件” 页面中
+- ⚠️ 将 “MCP” 页面中的工具管理移动到了 “插件” -> “管理行为” 中。
+- ⚠️ 将 “QQ 个人号（OneBot v11）” 机器人适配器类型更名为 “OneBot v11”，并将其 Logo 更改为 OneBot 的 Logo。
+- ⚠️ AstrBot WebChat 升级为 **AstrBot ChatUI**，入口从边栏修改为顶部（右上角）切换按钮。
+- 优化引用消息的逻辑，减少对模型输入缓存的破坏。
+- 优化当 Agent 达到最大步数时的处理。在达到最大步数后，会移除所有请求中的 tools 并告知模型根据上下文进行最终总结。
+- 优化 LLM tools 执行的错误处理，减少工具调用无限循环的问题。
+
+
+### 修复
+
+- ‼️ 修复部分情况下，分段回复无法正常分段的问题。
+- 修复处理工具返回结果的过程中，导致一些直接发送图片的工具（如生图工具）无法正确发送到用户的问题。
+- 修复 WebChat 部分情况下，上一条消息文字内容增量到下一条消息的问题。
+
+### 新增
+
+- 支持**指令管理**，设置指令别名、解决指令冲突、查看指令详情等。入口：“插件” -> “管理行为”。
+- 支持 Google Gemini 3 系列引入的 [Thinking Level](https://ai.google.dev/gemini-api/docs/thinking#thinking-levels) 配置。
+- 支持记录每条 LLM 消息的耗时、Token 使用量、TTFT 数据，以及每次 Agent Loop 的各种统计数据。
+- AstrBot ChatUI 支持查看每条消息的 TTFT、Token 使用量数据。
+- AstrBot ChatUI 支持显示每次工具调用的耗时、参数和响应。
+- AstrBot ChatUI 支持渲染 Mermaid、LateX 内容，优化了 Code Block 的显示效果（使用 Monaco Editor），并减少 DOM 更新于内存占用。（Powered by [Simon-He95/markstream-vue](https://github.com/Simon-He95/markstream-vue)）
+- 支持查看 Changelog 历史版本更新日志。
+- 🎄
+
+Merry Christmas! 

changelogs/v4.10.1.md

@@ -0,0 +1,46 @@
+## What's Changed
+
+> 📢 在升级前，请**完整阅读**本次更新日志。
+> 
+> **特别提醒：**
+> 1. 本次升级**如果再降级**，会由于提供商配置的变更，导致提供商配置错乱，需要手动删除后重新添加。
+> 2. 此版本 WebUI 包体相较上一个版本增加约 **193%**，共约 **9.8 MB**，升级可能会需要一些时间。
+> 3. **升级后请务必确保 WebUI 和 AstrBot Core 版本一致**，否则会产生预期之外的情况。（判断方法：日志中出现 `WebUI 版本已是最新。` 即为一致的版本，`检测到 WebUI 版本 (xxx) 与当前 AstrBot 版本 (xxx) 不符。` 即为不一致的版本。此版本的判断方法也可通查看 WebUI 右上角是否出现 Bot / Chat 的切换按钮控件来判断是否是新版本的 WebUI）。
+> 4. 如果有任何问题请提交 [Issue](https://github.com/AstrBotDevs/AstrBot/issues) 并附带 `v4.10.0` tag。
+
+## 4.10.0 -> 4.10.1
+
+- fix(core): 修复极少数情况下由于指令管理导致的 AstrBot 启动失败的问题
+- fix(core): 修复当提供商源带有斜杠（“/”）时，无法删除 / 更新提供商源的问题（报错 405）
+- perf(core): 优化 OneBot 适配器的消息段解析逻辑，修复部分情况下无法正确解析消息段的问题
+
+### 重构与优化
+
+- 重构 Provider 页面和提供商的配置结构，将 Chat Provider 配置拆分为 Provider Source（提供商源）和 Provider（代表提供商源的各个模型），引入了提供商模型自动发现、模型元数据自动发现的功能，**提供更加便捷的模型添加体验**。
+- ⚠️ 将 “MCP” 页面移动到了 “插件” 页面中
+- ⚠️ 将 “MCP” 页面中的工具管理移动到了 “插件” -> “管理行为” 中。
+- ⚠️ 将 “QQ 个人号（OneBot v11）” 机器人适配器类型更名为 “OneBot v11”，并将其 Logo 更改为 OneBot 的 Logo。
+- ⚠️ AstrBot WebChat 升级为 **AstrBot ChatUI**，入口从边栏修改为顶部（右上角）切换按钮。
+- 优化引用消息的逻辑，减少对模型输入缓存的破坏。
+- 优化当 Agent 达到最大步数时的处理。在达到最大步数后，会移除所有请求中的 tools 并告知模型根据上下文进行最终总结。
+- 优化 LLM tools 执行的错误处理，减少工具调用无限循环的问题。
+
+
+### 修复
+
+- ‼️ 修复部分情况下，分段回复无法正常分段的问题。
+- 修复处理工具返回结果的过程中，导致一些直接发送图片的工具（如生图工具）无法正确发送到用户的问题。
+- 修复 WebChat 部分情况下，上一条消息文字内容增量到下一条消息的问题。
+
+### 新增
+
+- 支持**指令管理**，设置指令别名、解决指令冲突、查看指令详情等。入口：“插件” -> “管理行为”。
+- 支持 Google Gemini 3 系列引入的 [Thinking Level](https://ai.google.dev/gemini-api/docs/thinking#thinking-levels) 配置。
+- 支持记录每条 LLM 消息的耗时、Token 使用量、TTFT 数据，以及每次 Agent Loop 的各种统计数据。
+- AstrBot ChatUI 支持查看每条消息的 TTFT、Token 使用量数据。
+- AstrBot ChatUI 支持显示每次工具调用的耗时、参数和响应。
+- AstrBot ChatUI 支持渲染 Mermaid、LateX 内容，优化了 Code Block 的显示效果（使用 Monaco Editor），并减少 DOM 更新于内存占用。（Powered by [Simon-He95/markstream-vue](https://github.com/Simon-He95/markstream-vue)）
+- 支持查看 Changelog 历史版本更新日志。
+- 🎄
+
+Merry Christmas! 

changelogs/v4.10.2.md

@@ -0,0 +1,9 @@
+## What's Changed
+
+### 修复
+
+1. ‼️‼️ 修复了由 `psutil` 新版本导致的启动时报错的问题。
+
+### 新增
+
+1. 插件指令管理支持管理别名。

dashboard/TAURI_README.md

@@ -1,225 +0,0 @@
-# AstrBot Dashboard - Tauri 桌面应用
-
-本项目现已支持通过 Tauri 构建为桌面应用，同时保持与 Web 版本的兼容性。
-
-## 环境要求
-
-### 系统依赖
-
-**macOS:**
-```bash
-# 安装 Xcode Command Line Tools
-xcode-select --install
-```
-
-**Windows:**
-- 安装 [Microsoft Visual Studio C++ Build Tools](https://visualstudio.microsoft.com/visual-cpp-build-tools/)
-- 安装 [WebView2](https://developer.microsoft.com/en-us/microsoft-edge/webview2/)
-
-**Linux (Ubuntu/Debian):**
-```bash
-sudo apt update
-sudo apt install libwebkit2gtk-4.0-dev \
-    build-essential \
-    curl \
-    wget \
-    file \
-    libssl-dev \
-    libgtk-3-dev \
-    libayatana-appindicator3-dev \
-    librsvg2-dev
-```
-
-### Rust 环境
-
-```bash
-# 安装 Rust
-curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
-
-# 验证安装
-rustc --version
-cargo --version
-```
-
-## 安装依赖
-
-```bash
-cd dashboard
-npm install
-```
-
-## 开发模式
-
-### Web 端开发（不变）
-
-```bash
-npm run dev
-```
-
-访问 http://localhost:3000
-
-### 桌面端开发
-
-```bash
-npm run tauri:dev
-```
-
-这会同时启动：
-1. Vite 开发服务器（端口 3000）
-2. Tauri 桌面应用窗口
-
-热重载功能正常工作，修改代码后会自动刷新。
-
-## 构建
-
-### Web 端构建（不变）
-
-```bash
-npm run build
-```
-
-输出目录：`dist/`
-
-### 桌面端构建
-
-```bash
-npm run tauri:build
-```
-
-构建产物位置：
-- **macOS**: `src-tauri/target/release/bundle/dmg/`
-- **Windows**: `src-tauri/target/release/bundle/msi/`
-- **Linux**: `src-tauri/target/release/bundle/deb/` 或 `appimage/`
-
-## 图标设置
-
-### 自动生成图标
-
-准备一个至少 512x512 像素的 PNG 图标，然后运行：
-
-```bash
-npm run tauri icon path/to/your/icon.png
-```
-
-### 手动设置图标
-
-将以下图标放入 `src-tauri/icons/` 目录：
-- `32x32.png`
-- `128x128.png`
-- `128x128@2x.png`
-- `icon.icns` (macOS)
-- `icon.ico` (Windows)
-
-## 代码兼容性
-
-项目已配置为同时支持 Web 和桌面端，使用相同的代码库。
-
-### 环境检测工具
-
-在 `src/utils/tauri.ts` 中提供了环境检测工具：
-
-```typescript
-import { isTauri, isWeb, PlatformAPI } from '@/utils/tauri';
-
-// 检测运行环境
-if (isTauri()) {
-  console.log('运行在桌面应用中');
-} else {
-  console.log('运行在浏览器中');
-}
-
-// 获取正确的 API 端点
-const baseURL = PlatformAPI.getBaseURL();
-```
-
-### API 调用注意事项
-
-- **Web 端**: 使用 Vite 代理，API 路径为 `/api/*`
-- **桌面端**: 直接连接到 `http://127.0.0.1:6185`
-
-已在 `PlatformAPI.getBaseURL()` 中处理，使用 axios 时：
-
-```typescript
-import axios from 'axios';
-import { PlatformAPI } from '@/utils/tauri';
-
-const api = axios.create({
-  baseURL: PlatformAPI.getBaseURL()
-});
-```
-
-## 配置说明
-
-### tauri.conf.json
-
-主要配置项：
-- `build.devPath`: 开发服务器地址（http://localhost:3000）
-- `build.distDir`: 构建输出目录（../dist）
-- `tauri.allowlist`: API 权限配置
-- `tauri.windows`: 窗口配置（大小、标题等）
-
-### 安全性
-
-默认配置已启用必要的权限：
-- 文件系统访问（限定在 APPDATA 目录）
-- HTTP 请求（限定到本地后端）
-- 窗口控制
-- 对话框（打开/保存文件）
-
-可在 `tauri.conf.json` 的 `allowlist` 部分调整权限。
-
-## 后端连接
-
-桌面应用需要后端服务运行在 `http://127.0.0.1:6185`。
-
-### 启动流程
-
-1. 启动 AstrBot 后端:
-   ```bash
-   cd /path/to/AstrBot
-   uv run main.py
-   ```
-
-2. 启动桌面应用:
-   ```bash
-   cd dashboard
-   npm run tauri:dev
-   ```
-
-或直接运行打包后的应用（后端需要已启动）。
-
-## 常见问题
-
-### Q: 桌面应用无法连接到后端？
-
-确保：
-1. AstrBot 后端正在运行（`uv run main.py`）
-2. 后端监听在 `127.0.0.1:6185`
-3. 防火墙未阻止连接
-
-### Q: 图标未显示？
-
-检查 `src-tauri/icons/` 目录中是否有所需的图标文件，或使用 `npm run tauri icon` 命令生成。
-
-### Q: 构建失败？
-
-- 确保已安装 Rust 和系统依赖
-- 运行 `cargo clean` 清理缓存后重试
-- 检查 Rust 版本（需要 1.60+）
-
-### Q: Web 端功能是否受影响？
-
-不受影响。`npm run dev` 和 `npm run build` 的行为完全不变。
-
-## 开发建议
-
-1. **优先使用 Web 端开发**: 更快的热重载，更好的调试体验
-2. **定期测试桌面端**: 确保跨平台兼容性
-3. **使用环境检测**: 针对不同平台提供最佳体验
-4. **注意 API 差异**: Web 和桌面端的某些 API 可能有差异
-
-## 更多资源
-
-- [Tauri 官方文档](https://tauri.app/)
-- [Tauri API 参考](https://tauri.app/v1/api/js/)
-- [Tauri Discord 社区](https://discord.com/invite/tauri)

dashboard/index.html

@@ -8,7 +8,7 @@
     <meta name="description" content="AstrBot Dashboard" />
     <link
       rel="stylesheet"
-      href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&family=Poppins:wght@400;500;600;700&family=Roboto:wght@400;500;700&display=swap"
+      href="https://fonts.googleapis.com/css2?family=Outfit&family=Poppins:wght@400;500;600;700&family=Roboto:wght@400;500;700&display=swap"
     />
     <title>AstrBot - 仪表盘</title>
   </head>

dashboard/package.json

@@ -10,30 +10,30 @@
     "build-prod": "vue-tsc --noEmit && vite build --base=/vue/free/",
     "preview": "vite preview --port 5050",
     "typecheck": "vue-tsc --noEmit",
-    "lint": "eslint . --ext .vue,.js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts --fix --ignore-path .gitignore",
+    "lint": "eslint . --ext .vue,.js,.jsx,.cjs,.mjs,.ts,.tsx,.cts,.mts --fix --ignore-path .gitignore"
-    "tauri": "tauri",
-    "tauri:dev": "tauri dev",
-    "tauri:build": "tauri build"
   },
   "dependencies": {
     "@guolao/vue-monaco-editor": "^1.5.4",
-    "@tauri-apps/api": "^2.9.0",
+    "@mdit/plugin-katex": "^0.24.1",
     "@tiptap/starter-kit": "2.1.7",
     "@tiptap/vue-3": "2.1.7",
     "apexcharts": "3.42.0",
     "axios": ">=1.6.2 <1.10.0 || >1.10.0 <2.0.0",
     "axios-mock-adapter": "^1.22.0",
     "chance": "1.1.11",
-    "d3": "^7.9.0",
     "date-fns": "2.30.0",
     "highlight.js": "^11.11.1",
     "js-md5": "^0.8.3",
+    "katex": "^0.16.27",
     "lodash": "4.17.21",
-    "marked": "^15.0.7",
+    "markstream-vue": "0.0.3-beta.7",
-    "markdown-it": "^14.1.0",
+    "mermaid": "^11.12.2",
-    "pinyin-pro": "^3.26.0",
     "pinia": "2.1.6",
+    "pinyin-pro": "^3.26.0",
     "remixicon": "3.5.0",
+    "shiki": "^3.20.0",
+    "stream-markdown": "^0.0.11",
+    "stream-monaco": "^0.0.8",
     "vee-validate": "4.11.3",
     "vite-plugin-vuetify": "1.0.2",
     "vue": "3.3.4",
@@ -47,9 +47,7 @@
   "devDependencies": {
     "@mdi/font": "7.2.96",
     "@rushstack/eslint-patch": "1.3.3",
-    "@tauri-apps/cli": "^2.9.4",
     "@types/chance": "1.1.3",
-    "@types/markdown-it": "^14.1.2",
     "@types/node": "^20.5.7",
     "@vitejs/plugin-vue": "4.3.3",
     "@vue/eslint-config-prettier": "8.0.0",

dashboard/src-tauri/.gitignore

@@ -1,3 +0,0 @@
-# Tauri specific
-src-tauri/target/
-src-tauri/WixTools/

dashboard/src-tauri/Cargo.toml

@@ -1,27 +0,0 @@
-[package]
-name = "astrbot-dashboard"
-version = "4.5.6"
-description = "AstrBot"
-authors = ["AstrBot Team"]
-license = "AGPL-3.0"
-repository = "https://github.com/AstrBotDevs/AstrBot"
-default-run = "astrbot-dashboard"
-edition = "2021"
-rust-version = "1.91.0"
-
-# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
-
-[build-dependencies]
-tauri-build = { version = "2", features = [] }
-
-[dependencies]
-serde_json = "1.0"
-serde = { version = "1.0", features = ["derive"] }
-tauri = { version = "2.9.2", features = ["macos-private-api", "protocol-asset"] }
-tauri-plugin-opener = "2"
-
-[features]
-# this feature is used for production builds or when `devPath` points to the filesystem and the built-in dev server is disabled.
-# If you use cargo directly instead of tauri's cli you can use this feature flag to switch between tauri's `dev` and `build` modes.
-# DO NOT REMOVE!!
-custom-protocol = [ "tauri/custom-protocol" ]

dashboard/src-tauri/build.rs

@@ -1,3 +0,0 @@
-fn main() {
-    tauri_build::build()
-}

dashboard/src-tauri/gen/schemas/capabilities.json

@@ -1 +0,0 @@
-{}

dashboard/src-tauri/icons/android/mipmap-anydpi-v26/ic_launcher.xml

@@ -1,5 +0,0 @@
-<?xml version="1.0" encoding="utf-8"?>
-<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
-  <foreground android:drawable="@mipmap/ic_launcher_foreground"/>
-  <background android:drawable="@color/ic_launcher_background"/>
-</adaptive-icon>

dashboard/src-tauri/icons/android/values/ic_launcher_background.xml

@@ -1,4 +0,0 @@
-<?xml version="1.0" encoding="utf-8"?>
-<resources>
-  <color name="ic_launcher_background">#fff</color>
-</resources>

dashboard/src-tauri/src/main.rs

@@ -1,104 +0,0 @@
-// Prevents additional console window on Windows in release, DO NOT REMOVE!!
-#![cfg_attr(not(debug_assertions), windows_subsystem = "windows")]
-
-use std::process::{Child, Command};
-use std::sync::Mutex;
-use tauri::{AppHandle, Emitter, Listener, Manager, State};
-
-struct BackendProcess(Mutex<Option<Child>>);
-
-fn start_backend_process(app_handle: &AppHandle) -> Option<Child> {
-    #[cfg(target_os = "macos")]
-    let backend_path = "astrbot-backend.app/Contents/MacOS/main";
-
-    #[cfg(target_os = "windows")]
-    let backend_path = "astrbot-backend.exe";
-
-    #[cfg(target_os = "linux")]
-    let backend_path = "astrbot-backend";
-
-    // 获取资源目录
-    let resource_dir = match app_handle
-        .path()
-        .resource_dir()
-    {
-        Ok(dir) => dir,
-        Err(e) => {
-            eprintln!("Failed to get resource directory: {}", e);
-            return None;
-        }
-    };
-
-    let full_backend_path = resource_dir.join(backend_path);
-
-    println!("Starting backend process at: {:?}", full_backend_path);
-
-    match Command::new(&full_backend_path).spawn() {
-        Ok(child) => {
-            println!(
-                "Backend process started successfully with PID: {}",
-                child.id()
-            );
-            Some(child)
-        }
-        Err(e) => {
-            eprintln!("Failed to start backend process: {}", e);
-            None
-        }
-    }
-}
-
-#[tauri::command]
-fn restart_backend(
-    app_handle: AppHandle,
-    backend_state: State<BackendProcess>,
-) -> Result<String, String> {
-    let mut backend = backend_state.0.lock().unwrap();
-
-    // 停止现有进程
-    if let Some(mut child) = backend.take() {
-        let _ = child.kill();
-        let _ = child.wait();
-    }
-
-    // 启动新进程
-    *backend = start_backend_process(&app_handle);
-
-    if backend.is_some() {
-        Ok("Backend restarted successfully".to_string())
-    } else {
-        Err("Failed to restart backend".to_string())
-    }
-}
-
-#[cfg_attr(mobile, tauri::mobile_entry_point)]
-pub fn run() {
-    tauri::Builder::default()
-        .setup(|app| {
-            // 启动后端进程
-            let backend_process = start_backend_process(app.handle());
-            app.manage(BackendProcess(Mutex::new(backend_process)));
-            Ok(())
-        })
-        .plugin(tauri_plugin_opener::init())
-        .invoke_handler(tauri::generate_handler![restart_backend])
-        .on_window_event(|window, event| {
-            if let tauri::WindowEvent::CloseRequested { .. } = event {
-                // 关闭窗口时清理后端进程
-                if let Some(backend_state) = window.app_handle().try_state::<BackendProcess>() {
-                    let mut backend = backend_state.0.lock().unwrap();
-                    if let Some(mut child) = backend.take() {
-                        let _ = child.kill();
-                        let _ = child.wait();
-                    }
-                }
-            }
-        })
-        .run(tauri::generate_context!())
-        .expect("error while running tauri application");
-}
-
-fn main() {
-    run();
-}
-

dashboard/src-tauri/tauri.conf.json

@@ -1,53 +0,0 @@
-{
-  "$schema": "https://schema.tauri.app/config/2",
-  "productName": "AstrBot",
-  "version": "4.5.6",
-  "identifier": "com.astrbot.app",
-  "build": {
-    "beforeDevCommand": "pnpm dev",
-    "devUrl": "http://localhost:3000",
-    "beforeBuildCommand": "pnpm build",
-    "frontendDist": "../dist"
-  },
-  "app": {
-    "withGlobalTauri": true,
-    "macOSPrivateApi": true,
-    "windows": [
-      {
-        "title": "AstrBot",
-        "label": "main",
-        "url": "/",
-        "width": 1400,
-        "height": 900
-      }
-    ],
-    "security": {
-      "csp": null,
-      "assetProtocol": {
-        "enable": true,
-        "scope": [
-          "$APPDATA/**"
-        ]
-      }
-    }
-  },
-  "bundle": {
-    "active": true,
-    "targets": "all",
-    "icon": [
-      "icons/32x32.png",
-      "icons/128x128.png",
-      "icons/128x128@2x.png",
-      "icons/icon.icns",
-      "icons/icon.ico"
-    ],
-    "resources": [
-      "resources/*"
-    ]
-  },
-  "plugins": {
-    "fs": {
-      "requireLiteralLeadingDot": false
-    }
-  }
-}