更新

2026-02-10 13:31:29 +08:00
38 changed files with 2956 additions and 282 deletions
--- a/Docs/BACKEND_DEV.md
+++ b/Docs/BACKEND_DEV.md
@@ -33,9 +33,10 @@ backend/
 │   │   ├── materials/       # 素材管理（router/schemas/service）
 │   │   ├── publish/         # 多平台发布
 │   │   ├── auth/            # 认证与会话
-│   │   ├── ai/              # AI 功能（标题标签生成等）
+│   │   ├── ai/              # AI 功能（标题标签生成、多语言翻译）
 │   │   ├── assets/          # 静态资源（字体/样式/BGM）
 │   │   ├── ref_audios/      # 声音克隆参考音频（router/schemas/service）
+│   │   ├── generated_audios/ # 预生成配音管理（router/schemas/service）
 │   │   ├── login_helper/    # 扫码登录辅助
 │   │   ├── tools/           # 工具接口（router/schemas/service）
 │   │   └── admin/           # 管理员功能
--- a/Docs/BACKEND_README.md
+++ b/Docs/BACKEND_README.md
@@ -19,12 +19,13 @@ backend/
 │   │   ├── materials/    # 素材管理（router/schemas/service）
 │   │   ├── publish/      # 多平台发布
 │   │   ├── auth/         # 认证与会话
-│   │   ├── ai/           # AI 功能（标题标签生成）
-│   │   ├── assets/       # 静态资源（字体/样式/BGM）
-│   │   ├── ref_audios/   # 声音克隆参考音频（router/schemas/service）
-│   │   ├── login_helper/ # 扫码登录辅助
-│   │   ├── tools/        # 工具接口（router/schemas/service）
-│   │   └── admin/        # 管理员功能
+│   │   ├── ai/              # AI 功能（标题标签生成、多语言翻译）
+│   │   ├── assets/          # 静态资源（字体/样式/BGM）
+│   │   ├── ref_audios/      # 声音克隆参考音频（router/schemas/service）
+│   │   ├── generated_audios/ # 预生成配音管理（router/schemas/service）
+│   │   ├── login_helper/    # 扫码登录辅助
+│   │   ├── tools/           # 工具接口（router/schemas/service）
+│   │   └── admin/           # 管理员功能
 │   ├── repositories/     # Supabase 数据访问
 │   ├── services/         # 外部服务集成 (TTS/Remotion/Storage/Uploader 等)
 │   └── tests/            # 单元测试与集成测试
@@ -83,11 +84,19 @@ backend/

 7.  **AI 功能 (AI)**
    *   `POST /api/ai/generate-meta`: AI 生成标题和标签
+    *   `POST /api/ai/translate`: AI 多语言翻译（支持 9 种目标语言）

-8.  **工具 (Tools)**
+8.  **预生成配音 (Generated Audios)**
+    *   `POST /api/generated-audios/generate`: 异步生成配音（返回 task_id）
+    *   `GET /api/generated-audios/tasks/{task_id}`: 轮询生成进度
+    *   `GET /api/generated-audios`: 列出用户所有配音
+    *   `DELETE /api/generated-audios/{audio_id}`: 删除配音
+    *   `PUT /api/generated-audios/{audio_id}`: 重命名配音
+
+9.  **工具 (Tools)**
    *   `POST /api/tools/extract-script`: 从视频链接提取文案

-9.  **健康检查**
+10. **健康检查**
    *   `GET /api/lipsync/health`: LatentSync 服务健康状态
    *   `GET /api/voiceclone/health`: Qwen3-TTS 服务健康状态

@@ -113,6 +122,9 @@ backend/
 - `tts_mode`: TTS 模式 (`edgetts` / `voiceclone`)
 - `voice`: EdgeTTS 音色 ID（edgetts 模式）
 - `ref_audio_id` / `ref_text`: 参考音频 ID 与文本（voiceclone 模式）
+- `generated_audio_id`: 预生成配音 ID（存在时跳过内联 TTS，使用已生成的配音文件）
+- `custom_assignments`: 自定义素材分配数组（每项含 `material_path` / `start` / `end` / `source_start`），存在时跳过 Whisper 均分
+- `language`: TTS 语言（默认自动检测，声音克隆时透传给 Qwen3-TTS）
 - `title`: 片头标题文字
 - `subtitle_style_id`: 字幕样式 ID
 - `title_style_id`: 标题样式 ID
--- a/Docs/DEPLOY_MANUAL.md
+++ b/Docs/DEPLOY_MANUAL.md
@@ -165,6 +165,8 @@ playwright install chromium
    CREATE POLICY "Allow public read" ON storage.objects FOR SELECT TO anon USING (bucket_id = 'materials' OR bucket_id = 'outputs');
    EOF
    ```
+
+> **注意**：后端启动时会自动创建额外的存储桶（`ref-audios`、`generated-audios`），无需手动创建。
 
 ---
 
@@ -570,6 +572,7 @@ pm2 logs vigent2-qwen-tts
 | `next` | React 框架 |
 | `swr` | 数据请求与缓存 |
 | `tailwindcss` | CSS 样式 |
+| `wavesurfer.js` | 音频波形（时间轴编辑器） |

 ### LatentSync 关键依赖

--- a/Docs/DevLogs/Day23.md
+++ b/Docs/DevLogs/Day23.md
@@ -0,0 +1,546 @@
+## 🎙️ 配音前置重构 — 第一阶段 (Day 23)
+
+### 概述
+
+将配音从视频生成流程中独立出来，实现"先生成配音 → 选中配音 → 再选素材 → 生成视频"的新工作流。用户可以独立管理配音（生成/试听/改名/删除/选择），并在选中配音后看到时长信息，为第二阶段的素材时间轴编排奠定数据基础。
+
+**旧流程**: 文案 + 选素材 → 一键生成（内联 TTS → Whisper → 均分 → LipSync → 合成）
+**新流程**: 文案 → 配音方式 → **生成配音** → 选中配音 → 选素材 → 背景音乐 → 生成视频
+
+---
+
+### 一、后端：新增 `generated_audios` 模块
+
+#### 模块结构
+
+```
+backend/app/modules/generated_audios/
+├── __init__.py
+├── router.py      # 5 个 API 端点
+├── schemas.py     # 请求/响应模型
+└── service.py     # 生成/列表/删除/改名
+```
+
+#### API 端点
+
+| 方法 | 路径 | 说明 |
+|------|------|------|
+| POST | `/api/generated-audios/generate` | 异步生成配音（返回 task_id） |
+| GET | `/api/generated-audios/tasks/{task_id}` | 轮询生成进度 |
+| GET | `/api/generated-audios` | 列出用户所有配音 |
+| DELETE | `/api/generated-audios/{audio_id}` | 删除配音 |
+| PUT | `/api/generated-audios/{audio_id}` | 改名 |
+
+#### 存储方案
+
+- Supabase 存储桶：`generated-audios`（启动时自动创建）
+- 音频文件：`{user_id}/{timestamp}_audio.wav`
+- 元数据文件：`{user_id}/{timestamp}_audio.json`（含 display_name、text、tts_mode、duration_sec 等）
+
+#### 生成流程
+
+复用现有 `TTSService` / `voice_clone_service` / `task_store`：
+
+```
+POST /generate → 创建 task → BackgroundTask:
+  1. edgetts → TTSService.generate_audio()
+     voiceclone → 下载 ref_audio → voice_clone_service.generate_audio()
+  2. ffprobe 获取时长
+  3. 上传 .wav + .json 到 generated-audios 桶
+  4. 更新 task(status=completed, output={audio_id, duration_sec, ...})
+```
+
+---
+
+### 二、后端：修改视频生成 workflow
+
+#### `GenerateRequest` 新增字段
+
+```python
+generated_audio_id: Optional[str] = None  # 预生成配音 ID（存在时跳过内联 TTS）
+```
+
+#### `workflow.py` TTS 阶段新增分支
+
+```python
+if req.generated_audio_id:
+    # 下载预生成配音 + 从元数据读取 language
+elif req.tts_mode == "voiceclone":
+    # 原有声音克隆逻辑
+else:
+    # 原有 EdgeTTS 逻辑
+```
+
+向后兼容：不传 `generated_audio_id` 时，原有内联 TTS 流程不受影响。
+
+---
+
+### 三、前端：新增配音列表 hook + 面板
+
+#### `useGeneratedAudios.ts`
+
+- 状态：`generatedAudios[]`、`selectedAudio`、`isGeneratingAudio`、`audioTask`
+- 方法：`fetchGeneratedAudios()`、`generateAudio()`、`deleteAudio()`、`renameAudio()`、`selectAudio()`
+- 轮询：生成后 1s 轮询 task 状态，完成后自动刷新列表并选中最新配音
+- 独立于视频生成的 TaskContext（不互相干扰）
+
+#### `GeneratedAudiosPanel.tsx`
+
+- 每条配音：播放/暂停、名称、时长、重命名、删除
+- 选中态：`border-purple-500 bg-purple-500/20`
+- 内嵌进度条（生成中显示）
+- 底部显示选中配音的原始文案（截断）
+- 播放逻辑自包含于面板内（`new Audio()` + play/pause toggle）
+
+---
+
+### 四、前端：UI 面板重排序
+
+**旧顺序**: MaterialSelector → ScriptEditor → TitleSubtitle → VoiceSelector → BgmPanel → GenerateActionBar
+
+**新顺序**:
+1. ScriptEditor（文案编辑）
+2. TitleSubtitlePanel（标题与字幕样式）
+3. VoiceSelector（配音方式）
+4. **GeneratedAudiosPanel**（配音列表）← 新增
+5. MaterialSelector（视频素材）← 后移，需选中配音才解锁
+6. BgmPanel（背景音乐）
+7. GenerateActionBar（生成视频）
+
+#### 素材区门控
+
+未选中配音时，素材区显示半透明遮罩 + "请先生成并选中配音"提示。素材上传/预览/改名/删除始终可用，仅选择勾选被遮罩。
+
+#### 时长信息
+
+选中配音后，MaterialSelector 顶部显示：
+```
+当前配音: 45.2 秒 | 已选 3 个素材（自动均分每段 ~15.1 秒）
+```
+
+#### 生成按钮条件更新
+
+```typescript
+// 旧条件
+disabled={isGenerating || selectedMaterials.length === 0 || (ttsMode === "voiceclone" && !selectedRefAudio)}
+// 新条件
+disabled={isGenerating || selectedMaterials.length === 0 || !selectedAudio}
+```
+
+---
+
+### 五、持久化
+
+`useHomePersistence` 新增 `selectedAudioId` 的 localStorage 读写，刷新页面后恢复选中的配音。
+
+---
+
+### 涉及文件汇总
+
+#### 后端新增
+
+| 文件 | 说明 |
+|------|------|
+| `backend/app/modules/generated_audios/__init__.py` | 模块标记 |
+| `backend/app/modules/generated_audios/router.py` | 5 个 API 端点 |
+| `backend/app/modules/generated_audios/service.py` | 生成/列表/删除/改名 |
+| `backend/app/modules/generated_audios/schemas.py` | 请求/响应模型 |
+
+#### 后端修改
+
+| 文件 | 变更 |
+|------|------|
+| `backend/app/main.py` | 注册 generated_audios 路由 |
+| `backend/app/services/storage.py` | 新增 `BUCKET_GENERATED_AUDIOS`，启动时自动创建桶 |
+| `backend/app/modules/videos/schemas.py` | `GenerateRequest` 新增 `generated_audio_id` 字段 |
+| `backend/app/modules/videos/workflow.py` | TTS 阶段新增预生成音频分支 |
+
+#### 前端新增
+
+| 文件 | 说明 |
+|------|------|
+| `frontend/src/features/home/model/useGeneratedAudios.ts` | 配音列表 hook |
+| `frontend/src/features/home/ui/GeneratedAudiosPanel.tsx` | 配音列表面板 |
+
+#### 前端修改
+
+| 文件 | 变更 |
+|------|------|
+| `frontend/src/features/home/ui/HomePage.tsx` | 面板重排序 + 素材区门控 + 插入 GeneratedAudiosPanel |
+| `frontend/src/features/home/ui/MaterialSelector.tsx` | 新增 `selectedAudioDuration` prop + 时长信息显示 |
+| `frontend/src/features/home/ui/GenerateActionBar.tsx` | 禁用条件改为 `!selectedAudio` |
+| `frontend/src/features/home/model/useHomeController.ts` | 集成 useGeneratedAudios、新增 handleGenerateAudio、修改 handleGenerate 使用 generated_audio_id |
+| `frontend/src/features/home/model/useHomePersistence.ts` | 新增 selectedAudioId 持久化 |
+
+---
+
+## 🎞️ 素材时间轴编排 — 第二阶段 (Day 23)
+
+### 概述
+
+在第一阶段"配音前置"基础上，新增**时间轴编辑器**，用户可以：
+1. 在音频波形上查看各素材块的时长分配
+2. 拖拽分割线调整每段素材的时长（无缝铺满，调整一段自动压缩/扩展相邻段）
+3. 为每段素材设置**源视频截取起点**（从视频任意位置开始，而非始终从头）
+
+**旧行为**: 多素材时自动均分（`_split_equal`），无法控制每段时长和源视频起始点
+**新行为**: 时间轴编辑器可视化分配 + 拖拽调整 + ClipTrimmer 截取设置
+
+---
+
+### 一、后端改动
+
+#### 1.1 新增 `CustomAssignment` 模型
+
+```python
+# backend/app/modules/videos/schemas.py
+class CustomAssignment(BaseModel):
+    material_path: str
+    start: float           # 音频时间轴起点
+    end: float             # 音频时间轴终点
+    source_start: float = 0.0  # 源视频截取起点
+```
+
+`GenerateRequest` 新增 `custom_assignments: Optional[List[CustomAssignment]] = None`。存在时跳过 Whisper 均分，直接使用用户定义的分配。
+
+#### 1.2 `prepare_segment` 支持 `source_start`
+
+```python
+def prepare_segment(self, video_path, target_duration, output_path,
+                    target_resolution=None, source_start: float = 0.0):
+```
+
+关键逻辑：
+- `source_start > 0` 时使用 `-ss` 快速 seek，并强制重编码（避免 stream copy 关键帧不精确）
+- 当需要循环且有 `source_start` 时，先裁剪出 `source_start` 到视频结尾的片段，再循环裁剪后的文件（避免 `stream_loop` 从视频 0s 开始循环）
+- 裁剪临时文件在 `finally` 中自动清理
+
+#### 1.3 `workflow.py` 支持 `custom_assignments`
+
+- **多素材模式**: `custom_assignments` 存在时，直接使用用户分配（仍运行 Whisper 生成字幕），每个 `prepare_segment` 调用传入 `source_start`
+- **单素材模式**: `custom_assignments` 有 1 条且 `source_start > 0` 时，先截取片段再传入 LatentSync
+- **向后兼容**: `custom_assignments` 为 `None` 时完全走旧路径
+
+---
+
+### 二、前端新增组件
+
+#### 2.1 `useTimelineEditor.ts` — 时间轴段管理 hook
+
+```typescript
+interface TimelineSegment {
+  id: string;              // React key
+  materialId: string;      // 素材 ID
+  materialName: string;    // 显示名
+  start: number;           // 音频时间轴开始秒数
+  end: number;             // 音频时间轴结束秒数
+  sourceStart: number;     // 源视频截取起点（默认 0）
+  sourceEnd: number;       // 源视频截取终点（0 = 到结尾）
+  color: string;           // 色块颜色
+}
+```
+
+核心方法：
+- `initSegments()`: selectedMaterials 变化时按数量均分 audioDuration
+- `resizeSegment(id, newEnd)`: 拖拽右边界，约束每段最小 1s
+- `setSourceRange(id, sourceStart, sourceEnd)`: 设置截取范围
+- `toCustomAssignments()`: 转为后端 `CustomAssignment[]` 格式
+
+#### 2.2 `TimelineEditor.tsx` — 波形 + 色块时间轴
+
+- **wavesurfer.js** 渲染音频波形（仅展示，不播放）
+- 色块层按比例排列，显示素材名 + 时长 + 截取标记
+- 色块间分割线可拖拽（`onPointerDown/Move/Up` 实现连续像素拖拽）
+- 点击色块打开 ClipTrimmer
+
+#### 2.3 `ClipTrimmer.tsx` — 素材截取模态框
+
+- HTML5 `<video>` 实时预览，拖拽滑块时 `video.currentTime` 跟随
+- 双端 Range Slider（起点/终点），互锁约束 ≥ 0.5s
+- 显示截取时长 vs 分配时长对比（循环补足/截断提示）
+- `loadedmetadata` 获取源视频时长
+
+---
+
+### 三、前端整合改动
+
+#### 3.1 `useHomeController.ts`
+
+- 集成 `useTimelineEditor` hook
+- 新增 `clipTrimmerOpen` / `clipTrimmerSegmentId` 状态
+- `handleGenerate` 多素材时始终发送 `custom_assignments`；单素材 + `sourceStart > 0` 时也发送
+- 移除不再使用的 `reorderMaterials` 导出
+
+#### 3.2 `HomePage.tsx`
+
+- 在 MaterialSelector 和 BgmPanel 之间插入 TimelineEditor（仅当有配音且已选素材时显示）
+- 底部新增 ClipTrimmer 模态框
+- 移除 `reorderMaterials` 和 `selectedAudioDuration` prop 传递
+
+#### 3.3 `MaterialSelector.tsx`
+
+- 移除配音时长信息栏（功能迁至 TimelineEditor）
+- 移除拖拽排序区（SortableChip + @dnd-kit 相关代码）
+- 移除 `onReorderMaterials` / `selectedAudioDuration` prop
+
+---
+
+### 四、审查修复的 Bug
+
+| # | 严重程度 | 问题 | 修复 |
+|---|---------|------|------|
+| 1 | **中** | `prepare_segment` 使用 `source_start > 0` + stream copy 时 seek 不精确 | 添加 `source_start > 0` 到重编码条件 |
+| 2 | **高** | `stream_loop + source_start` 循环时从视频 0s 开始而非从 source_start 循环 | 改为两步：先裁剪片段再循环裁剪后的文件 |
+| 3 | **低** | `useHomeController` 导出已废弃的 `reorderMaterials` | 移除 |
+
+---
+
+### 涉及文件汇总
+
+#### 后端修改
+
+| 文件 | 变更 |
+|------|------|
+| `backend/app/modules/videos/schemas.py` | 新增 `CustomAssignment` model，`GenerateRequest` 新增 `custom_assignments` 字段 |
+| `backend/app/services/video_service.py` | `prepare_segment` 新增 `source_start` 参数，循环+截取两步处理 |
+| `backend/app/modules/videos/workflow.py` | 多素材/单素材流水线支持 `custom_assignments`，传递 `source_start` |
+
+#### 前端新增
+
+| 文件 | 说明 |
+|------|------|
+| `frontend/src/features/home/model/useTimelineEditor.ts` | 时间轴段管理 hook |
+| `frontend/src/features/home/ui/TimelineEditor.tsx` | 波形 + 色块时间轴组件 |
+| `frontend/src/features/home/ui/ClipTrimmer.tsx` | 素材截取模态框 |
+
+#### 前端修改
+
+| 文件 | 变更 |
+|------|------|
+| `frontend/src/features/home/ui/HomePage.tsx` | 插入 TimelineEditor + ClipTrimmer |
+| `frontend/src/features/home/ui/MaterialSelector.tsx` | 移除时长信息 + 拖拽排序区 + 相关 prop |
+| `frontend/src/features/home/model/useHomeController.ts` | 集成 useTimelineEditor，handleGenerate 发送 custom_assignments |
+| `frontend/package.json` | 新增 `wavesurfer.js` 依赖 |
+
+---
+
+## 🎨 UI 体验优化 + TTS 稳定性修复 — 第三阶段 (Day 23)
+
+### 概述
+
+根据用户反馈，修复 6 项 UI 体验问题，同时修复 Qwen3-TTS 声音克隆服务的 SoX 路径问题和显存缓存管理。
+
+---
+
+### 一、Qwen3-TTS 稳定性修复
+
+#### 1.1 SoX PATH 修复
+
+**问题**: PM2 启动 qwen-tts 时，`sox` 工具安装在 conda env 的 bin 目录中，系统 PATH 找不到，导致音频编解码走 fallback 路径（CPU 密集型），日志中出现 `SoX could not be found!` 警告。
+
+**修复**: `run_qwen_tts.sh` 中 export conda env bin 到 PATH：
+
+```bash
+export PATH="/home/rongye/ProgramFiles/miniconda3/envs/qwen-tts/bin:$PATH"
+```
+
+#### 1.2 CUDA 缓存清理
+
+**修复**: `qwen_tts_server.py` 每次生成完成后（无论成功或失败）调用 `torch.cuda.empty_cache()`，防止显存碎片累积。使用 `asyncio.to_thread()` 在线程池中运行推理，避免阻塞事件循环导致健康检查超时。
+
+---
+
+### 二、配音列表按钮布局统一 (反馈 #1 + #6)
+
+**问题**: `GeneratedAudiosPanel` 的试听按钮位于左侧（独立于 Edit/Delete），与 `RefAudioPanel` 的布局不一致。底部文案摘要区域不需要展示。
+
+**修复**:
+- Play/Edit/Delete 按钮统一放在右侧同组，hover 显示，顺序为 试听→重命名→删除
+- 移除选中配音的文案摘要区域
+- 布局与 RefAudioPanel 一致：左侧名称+时长，右侧操作按钮组
+
+---
+
+### 三、视频素材区域移除配音依赖遮罩 (反馈 #2)
+
+**问题**: MaterialSelector 被 `!selectedAudio` 遮罩覆盖，必须先选配音才能操作素材。
+
+**修复**: 移除 `HomePage.tsx` 中 MaterialSelector 外层的 disabled overlay `<div>`。素材随时可上传/预览/管理，仅 TimelineEditor 需要选中配音才显示（已有独立条件 `selectedAudio && selectedMaterials.length > 0`）。
+
+---
+
+### 四、时间轴拖拽排序 (反馈 #3)
+
+**问题**: TimelineEditor 不支持调换素材顺序。
+
+**修复**:
+- `useTimelineEditor` 已有 `reorderSegments()` 方法（交换两个段的素材信息但保留时间范围）
+- 通过 `useHomeController` 暴露 `reorderSegments`，传入 `TimelineEditor`
+- 色块支持 HTML5 Drag & Drop：`draggable` + `onDragStart/Over/Drop/End`
+- 拖拽时：源色块半透明（`opacity-50`），目标色块高亮 ring（`ring-2 ring-purple-400 scale-[1.02]`）
+- 光标样式：`cursor-grab` / `active:cursor-grabbing`
+
+---
+
+### 五、截取设置双手柄 Range Slider (反馈 #4)
+
+**问题**: ClipTrimmer 使用两个独立的 `<input type="range">` 滑块，起点和终点分开操作，体验不直观。
+
+**修复**: 改为自定义双手柄 range slider：
+- 单条轨道，紫色圆形手柄（起点）+ 粉色圆形手柄（终点）
+- 轨道底色 `bg-white/10`，选中范围用素材对应颜色高亮
+- Pointer Events 实现拖拽：`onPointerDown` 捕获手柄 → `onPointerMove` 更新位置 → `onPointerUp` 释放
+- 手柄互锁约束：起点不超过终点 - 0.5s，终点不低于起点 + 0.5s
+- 底部显示起点（紫色）和终点（粉色）时间标签
+
+---
+
+### 六、截取设置视频预览 (反馈 #5)
+
+**问题**: ClipTrimmer 的视频只能静态查看，无法播放预览截取范围。
+
+**修复**:
+- 视频区域点击可播放/暂停（Play/Pause 图标覆盖层）
+- 播放范围：从 sourceStart 播放到 sourceEnd 自动停止
+- 播放结束后回到起点
+- 拖拽手柄时 `video.currentTime` 实时跟随（seek 到当前位置查看画面）
+- 播放进度条（白色竖线）叠加在 range slider 轨道上
+- `preload="auto"` 预加载视频，确保拖拽时快速 seek
+
+---
+
+### 涉及文件汇总
+
+#### 后端修改
+
+| 文件 | 变更 |
+|------|------|
+| `run_qwen_tts.sh` | export conda env bin 到 PATH，修复 SoX 找不到问题 |
+| `models/Qwen3-TTS/qwen_tts_server.py` | 每次生成后 `torch.cuda.empty_cache()`，asyncio.to_thread 避免阻塞 |
+
+#### 前端修改
+
+| 文件 | 变更 |
+|------|------|
+| `frontend/src/features/home/ui/GeneratedAudiosPanel.tsx` | 按钮布局统一（Play/Edit/Delete 右侧同组），移除文案摘要 |
+| `frontend/src/features/home/ui/HomePage.tsx` | 移除 MaterialSelector 配音遮罩，传入 onReorderSegment |
+| `frontend/src/features/home/ui/TimelineEditor.tsx` | 新增 HTML5 Drag & Drop 排序，新增 onReorderSegment prop |
+| `frontend/src/features/home/ui/ClipTrimmer.tsx` | 双手柄 range slider + 视频播放预览 + 播放进度指示 |
+| `frontend/src/features/home/model/useHomeController.ts` | 暴露 reorderSegments 方法 |
+
+---
+
+## 📝 历史文案保存 + 时间轴拖拽修复 — 第四阶段 (Day 23)
+
+### 概述
+
+新增文案手动保存与加载功能，修复时间轴拖拽排序后素材时长不跟随的 Bug，统一按钮视觉规范。
+
+---
+
+### 一、历史文案保存与加载
+
+#### 功能
+
+用户可手动保存当前文案到历史列表，随时从历史中加载恢复。只有手动保存的文案才出现在历史列表中，与自动保存（`useHomePersistence`）完全独立。
+
+#### UI 布局
+
+```
+按钮栏: [历史文案▼] [文案提取助手] [AI多语言▼] [AI生成标题标签]
+底部栏: 128 字                                    [保存文案]
+```
+
+- **历史文案下拉**: 展示已保存列表（名称 + 日期 + 删除按钮），点击条目加载文案，空列表显示"暂无保存的文案"
+- **保存文案按钮**: 文案为空时 disabled，点击后 `toast.success("文案已保存")`
+- **预计时长已移除**: 底部栏只保留字数 + 保存按钮
+
+#### 实现
+
+##### `useSavedScripts.ts`（新建）
+
+```typescript
+interface SavedScript { id: string; name: string; content: string; savedAt: number }
+```
+
+- localStorage key: `vigent_{storageKey}_savedScripts`
+- `saveScript(content)`: 取前 15 字符自动命名，新条目插入列表头部，**直接写入 localStorage**
+- `deleteScript(id)`: 删除指定条目，直接写入 localStorage
+- `useEffect([lsKey])`: lsKey 变化时（guest → userId）重新从 localStorage 读取
+- **不使用自动持久化 effect**，避免 storageKey 切换时空数组覆盖已有数据
+
+##### 数据流
+
+```
+ScriptEditor (UI)
+  ↑ savedScripts / onSaveScript / onLoadScript / onDeleteScript (纯 props + callbacks)
+  │
+useHomeController
+  ├── useSavedScripts(storageKey) → { savedScripts, saveScript, deleteScript }
+  └── handleSaveScript() → saveScript(text) + toast
+  │
+HomePage
+  └── 传递 props 到 ScriptEditor
+```
+
+---
+
+### 二、时间轴拖拽排序 Bug 修复
+
+#### 问题
+
+拖拽调换素材顺序后，各素材的时长没有跟随素材移动，而是留在原槽位。例如：素材1(3s) + 素材2(8s+4s循环)，拖拽后变成素材2(3s) + 素材1(8s+4s循环)，时长分配没变。
+
+#### 根因
+
+`reorderSegments` 使用**属性交换**方式：逐个拷贝 `materialId`、`sourceStart`、`sourceEnd` 等属性在两个槽位间交换，然后调用 `recalcPositions` 重算位置。
+
+#### 修复
+
+改为**数组移动**（splice）：将整个 segment 对象从旧位置取出，插入到新位置。segment 对象携带全部属性（materialId、sourceStart、sourceEnd、color 等）作为一个整体移动，再由 `recalcPositions` 重算位置。
+
+```typescript
+// 修复前：属性交换
+const fromMat = { materialId: next[fromIdx].materialId, ... };
+const toMat = { materialId: next[toIdx].materialId, ... };
+next[fromIdx] = { ...next[fromIdx], ...toMat };
+next[toIdx] = { ...next[toIdx], ...fromMat };
+
+// 修复后：数组移动
+const [moved] = next.splice(fromIdx, 1);
+next.splice(toIdx, 0, moved);
+```
+
+附带优势：3+ 素材拖拽行为从"交换"变为"插入"，更符合用户直觉。
+
+---
+
+### 三、按钮视觉统一
+
+#### 问题
+
+历史文案、文案提取助手、AI多语言、AI生成标题标签 4 个按钮高度不一致，AI 按钮的文本被 `<span>` 嵌套包裹导致内部布局差异。
+
+#### 修复
+
+- 4 个按钮统一为 `h-7 px-2.5 text-xs rounded inline-flex items-center gap-1`（固定高度 28px）
+- 移除 AI多语言 / AI生成标题标签 按钮内多余的 `<span>` 嵌套，改为 `<>...</>` fragment
+
+---
+
+### 涉及文件汇总
+
+#### 前端新增
+
+| 文件 | 说明 |
+|------|------|
+| `frontend/src/features/home/model/useSavedScripts.ts` | 历史文案 hook（localStorage 持久化） |
+
+#### 前端修改
+
+| 文件 | 变更 |
+|------|------|
+| `frontend/src/features/home/ui/ScriptEditor.tsx` | 历史文案下拉 + 保存按钮 + 移除预计时长 + 按钮高度统一 |
+| `frontend/src/features/home/model/useHomeController.ts` | 集成 useSavedScripts，新增 handleSaveScript |
+| `frontend/src/features/home/ui/HomePage.tsx` | 传递 savedScripts / handleSaveScript / deleteSavedScript 到 ScriptEditor |
+| `frontend/src/features/home/model/useTimelineEditor.ts` | reorderSegments 从属性交换改为数组移动（splice） |
--- a/Docs/FRONTEND_DEV.md
+++ b/Docs/FRONTEND_DEV.md
@@ -19,9 +19,12 @@ frontend/src/
 │   │   │   ├── useHomePersistence.ts   # 持久化管理
 │   │   │   ├── useBgm.ts
 │   │   │   ├── useGeneratedVideos.ts
+│   │   │   ├── useGeneratedAudios.ts
 │   │   │   ├── useMaterials.ts
 │   │   │   ├── useMediaPlayers.ts
 │   │   │   ├── useRefAudios.ts
+│   │   │   ├── useSavedScripts.ts
+│   │   │   ├── useTimelineEditor.ts
 │   │   │   └── useTitleSubtitleStyles.ts
 │   │   └── ui/                # UI 组件（纯 props + 回调）
 │   │       ├── HomePage.tsx
@@ -35,6 +38,9 @@ frontend/src/
 │   │       ├── FloatingStylePreview.tsx
 │   │       ├── VoiceSelector.tsx
 │   │       ├── RefAudioPanel.tsx
+│   │       ├── GeneratedAudiosPanel.tsx
+│   │       ├── TimelineEditor.tsx
+│   │       ├── ClipTrimmer.tsx
 │   │       ├── BgmPanel.tsx
 │   │       ├── GenerateActionBar.tsx
 │   │       ├── PreviewPanel.tsx
@@ -301,6 +307,15 @@ import { formatDate } from '@/shared/lib/media';
  - 标题字号 / 字幕字号
  - 背景音乐选择 / 音量 / 开关状态
  - 素材选择 / 历史作品选择
+  - 选中配音 ID (`selectedAudioId`)
+  - 时间轴段信息 (`useTimelineEditor` 的 localStorage)
+
+### 历史文案（独立持久化）
+
+`useSavedScripts` hook 独立管理历史文案的 localStorage 持久化：
+- key: `vigent_{storageKey}_savedScripts`
+- 仅在用户手动保存/删除时写入 localStorage，不使用自动持久化 effect
+- 与 `useHomePersistence` 完全独立，互不影响

 ### 实施规范
 - 使用 `storageKey = userId || 'guest'`，按用户隔离。
--- a/Docs/FRONTEND_README.md
+++ b/Docs/FRONTEND_README.md
@@ -17,7 +17,9 @@ ViGent2 的前端界面，采用 Next.js 16 + TailwindCSS 构建。
 - **作品预览**: 生成完成后直接播放下载（作品预览 + 历史作品）。
 - **预览优化**: 预览视频 `metadata` 预取，首帧加载更快。
 - **本地保存**: 文案/标题/偏好由 `useHomePersistence` 统一持久化，刷新后恢复 (Day 14/17)。
+- **历史文案**: 手动保存/加载/删除历史文案，独立 localStorage 持久化 (Day 23)。
 - **选择持久化**: 首页/发布页作品选择均使用稳定 `id` 持久化，刷新保持用户选择；新视频生成后自动选中最新 (Day 21)。
+- **AI 多语言翻译**: 支持 9 种目标语言翻译文案 + 还原原文 (Day 22)。

 ### 2. 全自动发布 (`/publish`) [Day 7 新增]
 - **多平台管理**: 统一管理抖音、微信视频号、B站、小红书账号状态。
@@ -35,8 +37,17 @@ ViGent2 的前端界面，采用 Next.js 16 + TailwindCSS 构建。
 - **TTS 模式选择**: EdgeTTS (预设音色) / 声音克隆 (自定义音色) 切换。
 - **参考音频管理**: 上传/列表/删除参考音频 (3-20秒 WAV)。
 - **一键克隆**: 选择参考音频后自动调用 Qwen3-TTS 服务。
+- **多语言支持**: EdgeTTS 10 语言声音列表，声音克隆 language 透传 (Day 22)。

-### 4. 字幕与标题 [Day 13 新增]
+### 4. 配音前置 + 时间轴编排 [Day 23 新增]
+- **配音独立生成**: 先生成配音 → 选中配音 → 再选素材 → 生成视频。
+- **配音管理面板**: 生成/试听/改名/删除/选中，异步生成 + 进度轮询。
+- **时间轴编辑器**: wavesurfer.js 音频波形 + 色块可视化素材分配，拖拽分割线调整各段时长。
+- **素材截取设置**: ClipTrimmer 双手柄 range slider + HTML5 视频预览播放。
+- **拖拽排序**: 时间轴色块支持 HTML5 Drag & Drop 调换素材顺序。
+- **自定义分配**: 后端 `custom_assignments` 支持用户定义的素材分配方案。
+
+### 5. 字幕与标题 [Day 13 新增]
 - **片头标题**: 可选输入，限制 15 字，视频开头显示 3 秒淡入淡出标题。
 - **标题同步**: 首页片头标题修改会同步到发布信息标题。
 - **逐字高亮字幕**: 卡拉OK效果，默认开启，可关闭。
@@ -45,16 +56,16 @@ ViGent2 的前端界面，采用 Next.js 16 + TailwindCSS 构建。
 - **默认样式**: 标题 90px 站酷快乐体；字幕 60px 经典黄字 + DingTalkJinBuTi (Day 17)。
 - **样式持久化**: 标题/字幕样式与字号刷新保留 (Day 17)。

-### 5. 背景音乐 [Day 16 新增]
+### 6. 背景音乐 [Day 16 新增]
 - **试听预览**: 点击试听即选中，音量滑块实时生效。
 - **混音控制**: 仅影响 BGM，配音保持原音量。

-### 6. 账户设置 [Day 15 新增]
+### 7. 账户设置 [Day 15 新增]
 - **手机号登录**: 11位中国手机号验证登录。
 - **账户下拉菜单**: 显示有效期 + 修改密码 + 安全退出。
 - **修改密码**: 弹窗输入当前密码与新密码，修改后强制重新登录。

-### 7. 文案提取助手 (`ScriptExtractionModal`) [Day 15 新增]
+### 8. 文案提取助手 (`ScriptExtractionModal`) [Day 15 新增]
 - **多源提取**: 支持文件拖拽上传与 URL 粘贴 (B站/抖音/TikTok)。
 - **AI 洗稿**: 集成 GLM-4.7-Flash，自动改写为口播文案。
 - **一键填入**: 提取结果直接填充至视频生成输入框。
@@ -66,6 +77,7 @@ ViGent2 的前端界面，采用 Next.js 16 + TailwindCSS 构建。
 - **样式**: TailwindCSS
 - **图标**: Lucide React
 - **组件**: 自定义现代化组件 (Glassmorphism 风格)
+- **音频波形**: wavesurfer.js (时间轴编辑器)
 - **API**: Axios 实例 `@/shared/api/axios` (对接后端 FastAPI :8006)

 ## 🚀 开发指南
--- a/Docs/QWEN3_TTS_DEPLOY.md
+++ b/Docs/QWEN3_TTS_DEPLOY.md
@@ -298,12 +298,20 @@ Response: audio/wav 文件
 SoX could not be found!
 ```

-**解决**: 通过 conda 安装 sox：
+**解决**:
+
+1. 通过 conda 安装 sox：

 ```bash
 conda install -y -c conda-forge sox
 ```

+2. 确保启动脚本 `run_qwen_tts.sh` 中已 export conda env bin 到 PATH（PM2 启动时系统 PATH 不含 conda 环境目录）：
+
+```bash
+export PATH="/home/rongye/ProgramFiles/miniconda3/envs/qwen-tts/bin:$PATH"
+```
+
 ### CUDA 内存不足

 Qwen3-TTS 1.7B 通常需要 8-10GB VRAM。如果遇到 OOM：
@@ -371,6 +379,7 @@ FOR INSERT TO anon WITH CHECK (bucket_id = 'ref-audios');

 | 日期 | 版本 | 说明 |
 |------|------|------|
+| 2026-02-09 | 1.2.0 | 修复 SoX PATH 问题（run_qwen_tts.sh export conda bin），每次生成后 empty_cache() |
 | 2026-01-30 | 1.1.0 | 明确默认模型升级为 1.7B-Base，替换旧版 0.6B 路径 |

 ---
--- a/Docs/SUBTITLE_DEPLOY.md
+++ b/Docs/SUBTITLE_DEPLOY.md
@@ -15,9 +15,13 @@
 原有流程:
  文本 → EdgeTTS → 音频 → LatentSync → FFmpeg合成 → 最终视频

-新流程:
-  文本 → EdgeTTS → 音频 ─┬→ LatentSync → 唇形视频 ─┐
-                        └→ faster-whisper → 字幕JSON ─┴→ Remotion合成 → 最终视频
+新流程 (单素材):
+  文本 → EdgeTTS/Qwen3-TTS/预生成配音 → 音频 ─┬→ LatentSync → 唇形视频 ─┐
+                                              └→ faster-whisper → 字幕JSON ─┴→ Remotion合成 → 最终视频
+
+新流程 (多素材):
+  音频 → 多素材按 custom_assignments 拼接 → LatentSync (单次推理) → 唇形视频 ─┐
+  音频 → faster-whisper → 字幕JSON ─────────────────────────────────────────────┴→ Remotion合成 → 最终视频
 ```

 ## 系统要求
@@ -140,7 +144,7 @@ remotion/
 | 阶段 | 进度 | 说明 |
 |------|------|------|
 | 下载素材 | 0% → 5% | 从 Supabase 下载输入视频 |
-| TTS 语音生成 | 5% → 25% | EdgeTTS 或 Qwen3-TTS 生成音频 |
+| TTS 语音生成 | 5% → 25% | EdgeTTS / Qwen3-TTS / 预生成配音下载 |
 | 唇形同步 | 25% → 80% | LatentSync 推理 |
 | 字幕对齐 | 80% → 85% | faster-whisper 生成字级别时间戳 |
 | Remotion 渲染 | 85% → 95% | 合成字幕和标题 |
@@ -282,4 +286,5 @@ WhisperService(device="cuda:0")  # 或 "cuda:1"
 | 日期 | 版本 | 说明 |
 |------|------|------|
 | 2026-01-29 | 1.0.0 | 初始版本，使用 faster-whisper + Remotion 实现逐字高亮字幕和片头标题 |
+| 2026-02-10 | 1.1.0 | 更新架构图：多素材 concat-then-infer、预生成配音选项 |
 | 2026-01-30 | 1.0.1 | 字幕高亮样式与标题动画优化，视觉表现更清晰 |
--- a/Docs/task_complete.md
+++ b/Docs/task_complete.md
@@ -1,8 +1,8 @@
 # ViGent2 开发任务清单 (Task Log)

 **项目**: ViGent2 数字人口播视频生成系统
-**进度**: 100% (Day 21 - 缺陷修复与持久化回归治理)
-**更新时间**: 2026-02-08
+**进度**: 100% (Day 23 - 配音前置重构 + 素材时间轴编排 + UI 体验优化)
+**更新时间**: 2026-02-10

 ---

@@ -10,7 +10,46 @@

 > 这里记录了每一天的核心开发内容与 milestone。

-### Day 21: 缺陷修复 + 浮动预览 + 发布重构 + 架构优化 + 多素材生成 (Current)
+### Day 23: 配音前置重构 + 素材时间轴编排 + UI 体验优化 + 历史文案 (Current)
+
+#### 第一阶段：配音前置
+- [x] **配音生成独立化**: 新增 `generated_audios` 后端模块（router/schemas/service），5 个 API 端点，复用现有 TTSService / voice_clone_service / task_store。
+- [x] **配音管理面板**: 前端新增 `useGeneratedAudios` hook + `GeneratedAudiosPanel` 组件，支持生成/试听/改名/删除/选中。
+- [x] **UI 面板重排序**: 文案 → 标题字幕 → 配音方式 → 配音列表 → 素材选择 → BGM → 生成视频。
+- [x] **素材区门控**: 未选中配音时素材区显示遮罩，选中后显示配音时长 + 素材均分信息。
+- [x] **视频生成对接**: workflow.py 新增预生成音频分支（`generated_audio_id`），跳过内联 TTS，向后兼容。
+- [x] **持久化**: selectedAudioId 加入 useHomePersistence，刷新页面恢复选中配音。
+
+#### 第二阶段：素材时间轴编排
+- [x] **时间轴编辑器**: 新增 `TimelineEditor` 组件，wavesurfer.js 音频波形 + 色块可视化素材分配，拖拽分割线调整各段时长。
+- [x] **素材截取设置**: 新增 `ClipTrimmer` 模态框，HTML5 视频预览 + 双端滑块设置源视频截取起点/终点。
+- [x] **后端自定义分配**: 新增 `CustomAssignment` 模型，`prepare_segment` 支持 `source_start`，workflow 多素材/单素材流水线支持 `custom_assignments`。
+- [x] **循环截取修复**: `stream_loop + source_start` 改为两步处理（先裁剪再循环），确保从截取起点循环而非从视频 0s 开始。
+- [x] **MaterialSelector 精简**: 移除旧的时长信息栏和拖拽排序区（功能迁移到 TimelineEditor）。
+
+#### 第三阶段：UI 体验优化 + TTS 稳定性
+- [x] **TTS SoX PATH 修复**: `run_qwen_tts.sh` export conda env bin 到 PATH，修复 `SoX could not be found!` 警告。
+- [x] **TTS 显存管理**: 每次生成后 `torch.cuda.empty_cache()`，asyncio.to_thread 避免阻塞事件循环。
+- [x] **配音列表按钮统一**: Play/Edit/Delete 按钮右侧同组 hover 显示，与 RefAudioPanel 一致，移除文案摘要。
+- [x] **素材区解除配音门控**: 移除 MaterialSelector 的 selectedAudio 遮罩，素材随时可上传管理。
+- [x] **时间轴拖拽排序**: TimelineEditor 色块支持 HTML5 Drag & Drop 调换素材顺序。
+- [x] **截取设置 Range Slider**: ClipTrimmer 改为单轨道双手柄（紫色起点+粉色终点），替换两个独立滑块。
+- [x] **截取设置视频预览**: 视频区域可播放/暂停，从 sourceStart 到 sourceEnd 自动停止，拖拽手柄时实时 seek。
+
+#### 第四阶段：历史文案 + Bug 修复
+- [x] **历史文案保存与加载**: 新增 `useSavedScripts` hook，手动保存/加载/删除历史文案，独立 localStorage 持久化。
+- [x] **时间轴拖拽修复**: `reorderSegments` 从属性交换改为数组移动（splice），修复拖拽后时长不跟随素材的 Bug。
+- [x] **按钮视觉统一**: 文案编辑区 4 个按钮统一为固定高度 `h-7`，移除多余 `<span>` 嵌套。
+- [x] **底部栏调整**: "保存文案"按钮移至底部右侧，移除预计时长显示。
+
+### Day 22: 多素材优化 + AI 翻译 + TTS 多语言
+- [x] **多素材 Bug 修复**: 6 个高优 Bug（边界溢出、单段 fallback、除零、duration 校验、Whisper 兜底、空列表检查）。
+- [x] **架构重构**: 多素材从"逐段 LatentSync"重构为"先拼接再推理"，推理次数 N→1。
+- [x] **前端优化**: payload 安全、进度消息、上传自动选中、Material 接口统一、拖拽修复、素材上限 4 个。
+- [x] **AI 多语言翻译**: 新增 `/api/ai/translate` 接口，前端 9 种语言翻译 + 还原原文。
+- [x] **TTS 多语言**: EdgeTTS 10 语言声音列表、翻译自动切换声音、声音克隆 language 透传、textLang 持久化。
+
+### Day 21: 缺陷修复 + 浮动预览 + 发布重构 + 架构优化 + 多素材生成
 - [x] **Remotion 崩溃容错**: 渲染进程 SIGABRT 退出时检查输出文件，避免误判失败导致标题/字幕丢失。
 - [x] **首页作品选择持久化**: 修复 `fetchGeneratedVideos` 无条件覆盖恢复值的问题，新增 `preferVideoId` 参数控制选中逻辑。
 - [x] **发布页作品选择持久化**: 根因为签名 URL 不稳定，全面改用 `video.id` 替代 `path` 进行选择/持久化/比较。
@@ -129,6 +168,7 @@
 ## 🛤️ 后续规划 (Roadmap)

 ### 🔴 优先待办
+- [x] ~~**配音前置重构 — 第二阶段**: 素材片段截取 + 语音时间轴编排~~ ✅ Day 23 已完成
 - [ ] **批量生成架构**: 支持 Excel 导入，批量生产视频。
 - [ ] **定时任务后台化**: 迁移前端触发的定时发布到后端 APScheduler。
 - [ ] **发布任务恢复机制**: 发布任务化 + 状态持久化 + 前端断点恢复，解决刷新后状态丢失。
@@ -146,7 +186,7 @@
 | **核心 API** | 100% | ✅ 稳定 |
 | **Web UI** | 100% | ✅ 稳定 (移动端适配) |
 | **唇形同步** | 100% | ✅ LatentSync 1.6 |
-| **TTS 配音** | 100% | ✅ EdgeTTS + Qwen3 |
+| **TTS 配音** | 100% | ✅ EdgeTTS + Qwen3 + 配音前置 + 时间轴编排 |
 | **自动发布** | 100% | ✅ 抖音/微信视频号/B站/小红书 |
 | **用户认证** | 100% | ✅ 手机号 + JWT |
 | **部署运维** | 100% | ✅ PM2 + Watchdog |
--- a/README.md
+++ b/README.md
@@ -17,13 +17,14 @@

 ### 核心能力
 - 🎬 **高清唇形同步** - LatentSync 1.6 驱动，512×512 高分辨率 Latent Diffusion 模型。
- 🎙️ **多模态配音** - 支持 **EdgeTTS** (微软超自然语音) 和 **Qwen3-TTS** (3秒极速声音克隆)。
+- 🎙️ **多模态配音** - 支持 **EdgeTTS** (微软超自然语音, 10 语言) 和 **Qwen3-TTS** (3秒极速声音克隆)。配音前置工作流：先生成配音 → 选素材 → 生成视频。
 - 📝 **智能字幕** - 集成 faster-whisper + Remotion，自动生成逐字高亮 (卡拉OK效果) 字幕。
 - 🎨 **样式预设** - 标题/字幕样式选择 + 预览 + 字号调节，支持自定义字体库。
 - 🖼️ **作品预览一致性** - 标题/字幕预览按素材分辨率缩放，效果更接近成片。
- 💾 **用户偏好持久化** - 首页状态统一恢复/保存，刷新后延续上次配置。
+- 🎞️ **多素材多机位** - 支持多选素材 + 时间轴编辑器 (wavesurfer.js 波形可视化)，拖拽分割线调整时长、拖拽排序切换机位、截取源视频片段。
+- 💾 **用户偏好持久化** - 首页状态统一恢复/保存，刷新后延续上次配置。历史文案手动保存与加载。
 - 🎵 **背景音乐** - 试听 + 音量控制 + 混音，保持配音音量稳定。
- 🤖 **AI 辅助创作** - 内置 GLM-4.7-Flash，支持 B站/抖音链接文案提取、AI 洗稿、标题/标签自动生成。
+- 🤖 **AI 辅助创作** - 内置 GLM-4.7-Flash，支持 B站/抖音链接文案提取、AI 洗稿、标题/标签自动生成、9 语言翻译。

 ### 平台化功能
 - 📱 **全自动发布** - 支持抖音/微信视频号/B站/小红书立即发布；扫码登录 + Cookie 持久化。
@@ -40,7 +41,7 @@

 | 领域 | 核心技术 | 说明 |
 |------|----------|------|
-| **前端** | Next.js 16 | TypeScript, TailwindCSS, SWR |
+| **前端** | Next.js 16 | TypeScript, TailwindCSS, SWR, wavesurfer.js |
 | **后端** | FastAPI | Python 3.10, AsyncIO, PM2 |
 | **数据库** | Supabase | PostgreSQL, Storage (本地/S3), Auth |
 | **唇形同步** | LatentSync 1.6 | PyTorch 2.5, Diffusers, DeepCache |
--- a/backend/app/main.py
+++ b/backend/app/main.py
@@ -15,6 +15,7 @@ from app.modules.ref_audios.router import router as ref_audios_router
 from app.modules.ai.router import router as ai_router
 from app.modules.tools.router import router as tools_router
 from app.modules.assets.router import router as assets_router
+from app.modules.generated_audios.router import router as generated_audios_router
 from loguru import logger
 import os

@@ -124,6 +125,7 @@ app.include_router(ref_audios_router, prefix="/api/ref-audios", tags=["RefAudios
 app.include_router(ai_router)  # /api/ai
 app.include_router(tools_router, prefix="/api/tools", tags=["Tools"])
 app.include_router(assets_router, prefix="/api/assets", tags=["Assets"])
+app.include_router(generated_audios_router, prefix="/api/generated-audios", tags=["GeneratedAudios"])


@app.on_event("startup")
--- a/backend/app/modules/generated_audios/init.py
+++ b/backend/app/modules/generated_audios/init.py
--- a/backend/app/modules/generated_audios/router.py
+++ b/backend/app/modules/generated_audios/router.py
@@ -0,0 +1,77 @@
+"""生成配音 API"""
+from fastapi import APIRouter, BackgroundTasks, Depends, HTTPException
+import uuid
+from loguru import logger
+
+from app.core.deps import get_current_user
+from app.core.response import success_response
+from app.modules.videos.task_store import create_task, get_task
+from app.modules.generated_audios.schemas import GenerateAudioRequest, RenameAudioRequest
+from app.modules.generated_audios import service
+
+router = APIRouter()
+
+
+@router.post("/generate")
+async def generate_audio(
+    req: GenerateAudioRequest,
+    background_tasks: BackgroundTasks,
+    user: dict = Depends(get_current_user),
+):
+    """异步生成配音（返回 task_id）"""
+    task_id = str(uuid.uuid4())
+    create_task(task_id, user["id"])
+    background_tasks.add_task(service.generate_audio_task, task_id, req, user["id"])
+    return success_response({"task_id": task_id})
+
+
+@router.get("/tasks/{task_id}")
+async def get_audio_task(task_id: str, user: dict = Depends(get_current_user)):
+    """轮询配音生成进度"""
+    task = get_task(task_id)
+    if task.get("status") != "not_found" and task.get("user_id") != user["id"]:
+        return success_response({"status": "not_found"})
+    return success_response(task)
+
+
+@router.get("")
+async def list_audios(user: dict = Depends(get_current_user)):
+    """列出当前用户所有已生成配音"""
+    try:
+        result = await service.list_generated_audios(user["id"])
+        return success_response(result)
+    except Exception as e:
+        logger.error(f"列出配音失败: {e}")
+        raise HTTPException(status_code=500, detail=f"获取列表失败: {str(e)}")
+
+
+@router.delete("/{audio_id:path}")
+async def delete_audio(audio_id: str, user: dict = Depends(get_current_user)):
+    """删除配音"""
+    try:
+        await service.delete_generated_audio(audio_id, user["id"])
+        return success_response(message="删除成功")
+    except PermissionError as e:
+        raise HTTPException(status_code=403, detail=str(e))
+    except Exception as e:
+        logger.error(f"删除配音失败: {e}")
+        raise HTTPException(status_code=500, detail=f"删除失败: {str(e)}")
+
+
+@router.put("/{audio_id:path}")
+async def rename_audio(
+    audio_id: str,
+    request: RenameAudioRequest,
+    user: dict = Depends(get_current_user),
+):
+    """重命名配音"""
+    try:
+        result = await service.rename_generated_audio(audio_id, request.new_name, user["id"])
+        return success_response(result, message="重命名成功")
+    except PermissionError as e:
+        raise HTTPException(status_code=403, detail=str(e))
+    except ValueError as e:
+        raise HTTPException(status_code=400, detail=str(e))
+    except Exception as e:
+        logger.error(f"重命名配音失败: {e}")
+        raise HTTPException(status_code=500, detail=f"重命名失败: {str(e)}")
--- a/backend/app/modules/generated_audios/schemas.py
+++ b/backend/app/modules/generated_audios/schemas.py
@@ -0,0 +1,30 @@
+from pydantic import BaseModel
+from typing import Optional, List
+
+
+class GenerateAudioRequest(BaseModel):
+    text: str
+    tts_mode: str = "edgetts"
+    voice: str = "zh-CN-YunxiNeural"
+    ref_audio_id: Optional[str] = None
+    ref_text: Optional[str] = None
+    language: str = "zh-CN"
+
+
+class RenameAudioRequest(BaseModel):
+    new_name: str
+
+
+class GeneratedAudioItem(BaseModel):
+    id: str
+    name: str
+    path: str
+    duration_sec: float
+    text: str
+    tts_mode: str
+    language: str
+    created_at: int
+
+
+class GeneratedAudioListResponse(BaseModel):
+    items: List[GeneratedAudioItem]
--- a/backend/app/modules/generated_audios/service.py
+++ b/backend/app/modules/generated_audios/service.py
@@ -0,0 +1,263 @@
+"""生成配音 - 业务逻辑"""
+import re
+import json
+import time
+import asyncio
+import subprocess
+import tempfile
+import os
+from pathlib import Path
+from typing import Optional
+
+import httpx
+from loguru import logger
+
+from app.services.storage import storage_service
+from app.services.tts_service import TTSService
+from app.services.voice_clone_service import voice_clone_service
+from app.modules.videos.task_store import task_store
+from app.modules.generated_audios.schemas import (
+    GenerateAudioRequest,
+    GeneratedAudioItem,
+    GeneratedAudioListResponse,
+)
+
+BUCKET = "generated-audios"
+
+
+def _locale_to_qwen_lang(locale: str) -> str:
+    mapping = {"zh": "Chinese", "en": "English"}
+    return mapping.get(locale.split("-")[0], "Auto")
+
+
+def _get_audio_duration(file_path: str) -> float:
+    try:
+        result = subprocess.run(
+            ['ffprobe', '-v', 'quiet', '-show_entries', 'format=duration',
+             '-of', 'csv=p=0', file_path],
+            capture_output=True, text=True, timeout=10
+        )
+        return float(result.stdout.strip())
+    except Exception as e:
+        logger.warning(f"获取音频时长失败: {e}")
+        return 0.0
+
+
+async def generate_audio_task(task_id: str, req: GenerateAudioRequest, user_id: str):
+    """后台任务：生成配音"""
+    try:
+        task_store.update(task_id, {"status": "processing", "progress": 10, "message": "正在生成配音..."})
+
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp:
+            audio_path = tmp.name
+
+        try:
+            if req.tts_mode == "voiceclone":
+                if not req.ref_audio_id or not req.ref_text:
+                    raise ValueError("声音克隆模式需要提供参考音频和参考文字")
+
+                task_store.update(task_id, {"progress": 20, "message": "正在下载参考音频..."})
+
+                with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as tmp_ref:
+                    ref_local = tmp_ref.name
+
+                try:
+                    ref_url = await storage_service.get_signed_url(
+                        bucket="ref-audios", path=req.ref_audio_id
+                    )
+                    timeout = httpx.Timeout(None)
+                    async with httpx.AsyncClient(timeout=timeout) as client:
+                        async with client.stream("GET", ref_url) as resp:
+                            resp.raise_for_status()
+                            with open(ref_local, "wb") as f:
+                                async for chunk in resp.aiter_bytes():
+                                    f.write(chunk)
+
+                    task_store.update(task_id, {"progress": 40, "message": "正在克隆声音 (Qwen3-TTS)..."})
+                    await voice_clone_service.generate_audio(
+                        text=req.text,
+                        ref_audio_path=ref_local,
+                        ref_text=req.ref_text,
+                        output_path=audio_path,
+                        language=_locale_to_qwen_lang(req.language),
+                    )
+                finally:
+                    if os.path.exists(ref_local):
+                        os.unlink(ref_local)
+            else:
+                task_store.update(task_id, {"progress": 30, "message": "正在生成语音 (EdgeTTS)..."})
+                tts = TTSService()
+                await tts.generate_audio(req.text, req.voice, audio_path)
+
+            task_store.update(task_id, {"progress": 70, "message": "正在上传配音..."})
+
+            duration = _get_audio_duration(audio_path)
+            timestamp = int(time.time())
+            audio_id = f"{user_id}/{timestamp}_audio.wav"
+            meta_id = f"{user_id}/{timestamp}_audio.json"
+
+            # 生成 display_name
+            now = time.strftime("%Y%m%d_%H%M", time.localtime(timestamp))
+            display_name = f"配音_{now}"
+
+            with open(audio_path, "rb") as f:
+                wav_data = f.read()
+
+            await storage_service.upload_file(
+                bucket=BUCKET, path=audio_id,
+                file_data=wav_data, content_type="audio/wav",
+            )
+
+            metadata = {
+                "display_name": display_name,
+                "text": req.text,
+                "tts_mode": req.tts_mode,
+                "voice": req.voice if req.tts_mode == "edgetts" else None,
+                "ref_audio_id": req.ref_audio_id,
+                "language": req.language,
+                "duration_sec": duration,
+                "created_at": timestamp,
+            }
+            await storage_service.upload_file(
+                bucket=BUCKET, path=meta_id,
+                file_data=json.dumps(metadata, ensure_ascii=False).encode("utf-8"),
+                content_type="application/json",
+            )
+
+            signed_url = await storage_service.get_signed_url(BUCKET, audio_id)
+
+            task_store.update(task_id, {
+                "status": "completed",
+                "progress": 100,
+                "message": f"配音生成完成 ({duration:.1f}s)",
+                "output": {
+                    "audio_id": audio_id,
+                    "name": display_name,
+                    "path": signed_url,
+                    "duration_sec": duration,
+                    "text": req.text,
+                    "tts_mode": req.tts_mode,
+                    "language": req.language,
+                    "created_at": timestamp,
+                },
+            })
+        finally:
+            if os.path.exists(audio_path):
+                os.unlink(audio_path)
+
+    except Exception as e:
+        import traceback
+        task_store.update(task_id, {
+            "status": "failed",
+            "message": f"配音生成失败: {str(e)}",
+            "error": traceback.format_exc(),
+        })
+        logger.error(f"Generate audio failed: {e}")
+
+
+async def list_generated_audios(user_id: str) -> dict:
+    """列出用户的所有已生成配音"""
+    files = await storage_service.list_files(BUCKET, user_id)
+    wav_files = [f for f in files if f.get("name", "").endswith("_audio.wav")]
+
+    if not wav_files:
+        return GeneratedAudioListResponse(items=[]).model_dump()
+
+    async def fetch_info(f):
+        name = f.get("name", "")
+        storage_path = f"{user_id}/{name}"
+        meta_name = name.replace("_audio.wav", "_audio.json")
+        meta_path = f"{user_id}/{meta_name}"
+
+        display_name = name
+        text = ""
+        tts_mode = "edgetts"
+        language = "zh-CN"
+        duration_sec = 0.0
+        created_at = 0
+
+        try:
+            meta_url = await storage_service.get_signed_url(BUCKET, meta_path)
+            async with httpx.AsyncClient(timeout=5.0) as client:
+                resp = await client.get(meta_url)
+                if resp.status_code == 200:
+                    meta = resp.json()
+                    display_name = meta.get("display_name", name)
+                    text = meta.get("text", "")
+                    tts_mode = meta.get("tts_mode", "edgetts")
+                    language = meta.get("language", "zh-CN")
+                    duration_sec = meta.get("duration_sec", 0.0)
+                    created_at = meta.get("created_at", 0)
+        except Exception as e:
+            logger.debug(f"读取配音 metadata 失败: {e}")
+            try:
+                created_at = int(name.split("_")[0])
+            except:
+                pass
+
+        signed_url = await storage_service.get_signed_url(BUCKET, storage_path)
+
+        return GeneratedAudioItem(
+            id=storage_path,
+            name=display_name,
+            path=signed_url,
+            duration_sec=duration_sec,
+            text=text,
+            tts_mode=tts_mode,
+            language=language,
+            created_at=created_at,
+        )
+
+    items = await asyncio.gather(*[fetch_info(f) for f in wav_files])
+    items = sorted(items, key=lambda x: x.created_at, reverse=True)
+    return GeneratedAudioListResponse(items=items).model_dump()
+
+
+async def delete_generated_audio(audio_id: str, user_id: str) -> None:
+    if not audio_id.startswith(f"{user_id}/"):
+        raise PermissionError("无权删除此文件")
+
+    await storage_service.delete_file(BUCKET, audio_id)
+    meta_path = audio_id.replace("_audio.wav", "_audio.json")
+    try:
+        await storage_service.delete_file(BUCKET, meta_path)
+    except:
+        pass
+
+
+async def rename_generated_audio(audio_id: str, new_name: str, user_id: str) -> dict:
+    if not audio_id.startswith(f"{user_id}/"):
+        raise PermissionError("无权修改此文件")
+
+    new_name = new_name.strip()
+    if not new_name:
+        raise ValueError("新名称不能为空")
+
+    meta_path = audio_id.replace("_audio.wav", "_audio.json")
+    try:
+        meta_url = await storage_service.get_signed_url(BUCKET, meta_path)
+        async with httpx.AsyncClient() as client:
+            resp = await client.get(meta_url)
+            if resp.status_code == 200:
+                metadata = resp.json()
+            else:
+                raise Exception(f"Failed to fetch metadata: {resp.status_code}")
+    except Exception as e:
+        logger.warning(f"无法读取配音元数据: {e}, 将创建新的")
+        metadata = {
+            "display_name": new_name,
+            "text": "",
+            "tts_mode": "edgetts",
+            "language": "zh-CN",
+            "duration_sec": 0.0,
+            "created_at": int(time.time()),
+        }
+
+    metadata["display_name"] = new_name
+    await storage_service.upload_file(
+        bucket=BUCKET,
+        path=meta_path,
+        file_data=json.dumps(metadata, ensure_ascii=False).encode("utf-8"),
+        content_type="application/json",
+    )
+    return {"name": new_name}
--- a/backend/app/modules/videos/schemas.py
+++ b/backend/app/modules/videos/schemas.py
@@ -2,6 +2,13 @@ from pydantic import BaseModel
 from typing import Optional, List


+class CustomAssignment(BaseModel):
+    material_path: str
+    start: float           # 音频时间轴起点
+    end: float             # 音频时间轴终点
+    source_start: float = 0.0  # 源视频截取起点
+
+
 class GenerateRequest(BaseModel):
    text: str
    voice: str = "zh-CN-YunxiNeural"
@@ -11,6 +18,7 @@ class GenerateRequest(BaseModel):
    ref_audio_id: Optional[str] = None
    ref_text: Optional[str] = None
    language: str = "zh-CN"
+    generated_audio_id: Optional[str] = None  # 预生成配音 ID（存在时跳过内联 TTS）
    title: Optional[str] = None
    enable_subtitles: bool = True
    subtitle_style_id: Optional[str] = None
@@ -21,3 +29,4 @@ class GenerateRequest(BaseModel):
    subtitle_bottom_margin: Optional[int] = None
    bgm_id: Optional[str] = None
    bgm_volume: Optional[float] = 0.2
+    custom_assignments: Optional[List[CustomAssignment]] = None
--- a/backend/app/modules/videos/workflow.py
+++ b/backend/app/modules/videos/workflow.py
@@ -197,7 +197,33 @@ async def process_video_generation(task_id: str, req: GenerateRequest, user_id:
        audio_path = temp_dir / f"{task_id}_audio.wav"
        temp_files.append(audio_path)

-        if req.tts_mode == "voiceclone":
+        if req.generated_audio_id:
+            # 新流程：使用预生成的配音
+            _update_task(task_id, message="正在下载配音...", progress=12)
+            audio_url = await storage_service.get_signed_url(
+                bucket="generated-audios",
+                path=req.generated_audio_id,
+            )
+            await _download_material(audio_url, audio_path)
+
+            # 从元数据获取 language
+            meta_path = req.generated_audio_id.replace("_audio.wav", "_audio.json")
+            try:
+                meta_url = await storage_service.get_signed_url(
+                    bucket="generated-audios", path=meta_path,
+                )
+                import httpx as _httpx
+                async with _httpx.AsyncClient(timeout=5.0) as client:
+                    resp = await client.get(meta_url)
+                    if resp.status_code == 200:
+                        meta = resp.json()
+                        req.language = meta.get("language", req.language)
+                        if not req.text.strip():
+                            req.text = meta.get("text", req.text)
+            except Exception as e:
+                logger.warning(f"读取配音元数据失败: {e}")
+
+        elif req.tts_mode == "voiceclone":
            if not req.ref_audio_id or not req.ref_text:
                raise ValueError("声音克隆模式需要提供参考音频和参考文字")

@@ -239,40 +265,74 @@ async def process_video_generation(task_id: str, req: GenerateRequest, user_id:
            # ══════════════════════════════════════
            # 多素材流水线
            # ══════════════════════════════════════
-            _update_task(task_id, progress=12, message="正在生成字幕 (Whisper)...")
+            _update_task(task_id, progress=12, message="正在分配素材...")

-            captions_path = temp_dir / f"{task_id}_captions.json"
-            temp_files.append(captions_path)
-
-            try:
-                captions_data = await whisper_service.align(
-                    audio_path=str(audio_path),
-                    text=req.text,
-                    output_path=str(captions_path),
-                    language=_locale_to_whisper_lang(req.language),
-                )
-                print(f"[Pipeline] Whisper alignment completed (multi-material)")
-            except Exception as e:
-                logger.warning(f"Whisper alignment failed: {e}")
-                captions_data = None
-                captions_path = None
-
-            _update_task(task_id, progress=15, message="正在分配素材...")
-
-            if captions_data and captions_data.get("segments"):
-                assignments = _split_equal(captions_data["segments"], material_paths)
-            else:
-                # Whisper 失败 → 按时长均分（不依赖字符对齐）
-                logger.warning("[MultiMat] Whisper 无数据，按时长均分")
-                audio_dur = video._get_duration(str(audio_path))
-                if audio_dur <= 0:
-                    audio_dur = 30.0  # 安全兜底
-                seg_dur = audio_dur / len(material_paths)
+            if req.custom_assignments:
+                # 用户自定义分配，跳过 Whisper 均分
                assignments = [
-                    {"material_path": material_paths[i], "start": i * seg_dur,
-                     "end": (i + 1) * seg_dur, "index": i}
-                    for i in range(len(material_paths))
+                    {
+                        "material_path": a.material_path,
+                        "start": a.start,
+                        "end": a.end,
+                        "source_start": a.source_start,
+                        "index": i,
+                    }
+                    for i, a in enumerate(req.custom_assignments)
                ]
+                # 仍然需要 Whisper 生成字幕（如果启用）
+                captions_path = temp_dir / f"{task_id}_captions.json"
+                temp_files.append(captions_path)
+                if req.enable_subtitles:
+                    _update_task(task_id, message="正在生成字幕 (Whisper)...")
+                    try:
+                        await whisper_service.align(
+                            audio_path=str(audio_path),
+                            text=req.text,
+                            output_path=str(captions_path),
+                            language=_locale_to_whisper_lang(req.language),
+                        )
+                        print(f"[Pipeline] Whisper alignment completed (custom assignments)")
+                    except Exception as e:
+                        logger.warning(f"Whisper alignment failed: {e}")
+                        captions_path = None
+                else:
+                    captions_path = None
+            else:
+                # 原有逻辑：Whisper → _split_equal
+                _update_task(task_id, message="正在生成字幕 (Whisper)...")
+
+                captions_path = temp_dir / f"{task_id}_captions.json"
+                temp_files.append(captions_path)
+
+                try:
+                    captions_data = await whisper_service.align(
+                        audio_path=str(audio_path),
+                        text=req.text,
+                        output_path=str(captions_path),
+                        language=_locale_to_whisper_lang(req.language),
+                    )
+                    print(f"[Pipeline] Whisper alignment completed (multi-material)")
+                except Exception as e:
+                    logger.warning(f"Whisper alignment failed: {e}")
+                    captions_data = None
+                    captions_path = None
+
+                _update_task(task_id, progress=15, message="正在分配素材...")
+
+                if captions_data and captions_data.get("segments"):
+                    assignments = _split_equal(captions_data["segments"], material_paths)
+                else:
+                    # Whisper 失败 → 按时长均分（不依赖字符对齐）
+                    logger.warning("[MultiMat] Whisper 无数据，按时长均分")
+                    audio_dur = video._get_duration(str(audio_path))
+                    if audio_dur <= 0:
+                        audio_dur = 30.0  # 安全兜底
+                    seg_dur = audio_dur / len(material_paths)
+                    assignments = [
+                        {"material_path": material_paths[i], "start": i * seg_dur,
+                         "end": (i + 1) * seg_dur, "index": i}
+                        for i in range(len(material_paths))
+                    ]

            # 扩展段覆盖完整音频范围：首段从0开始，末段到音频结尾
            audio_duration = video._get_duration(str(audio_path))
@@ -321,7 +381,8 @@ async def process_video_generation(task_id: str, req: GenerateRequest, user_id:
                temp_files.append(prepared_path)
                video.prepare_segment(
                    str(material_locals[i]), seg_dur, str(prepared_path),
-                    target_resolution=base_res if need_scale else None
+                    target_resolution=base_res if need_scale else None,
+                    source_start=assignment.get("source_start", 0.0),
                )
                prepared_segments.append(prepared_path)

@@ -363,6 +424,25 @@ async def process_video_generation(task_id: str, req: GenerateRequest, user_id:
            # ══════════════════════════════════════
            # 单素材流水线（原有逻辑）
            # ══════════════════════════════════════
+
+            # 单素材 + source_start：先截取片段
+            single_source_start = 0.0
+            if req.custom_assignments and len(req.custom_assignments) == 1:
+                single_source_start = req.custom_assignments[0].source_start
+
+            if single_source_start > 0:
+                _update_task(task_id, progress=20, message="正在截取素材片段...")
+                audio_dur = video._get_duration(str(audio_path))
+                if audio_dur <= 0:
+                    audio_dur = 30.0
+                trimmed_path = temp_dir / f"{task_id}_trimmed.mp4"
+                temp_files.append(trimmed_path)
+                video.prepare_segment(
+                    str(input_material_path), audio_dur, str(trimmed_path),
+                    source_start=single_source_start,
+                )
+                input_material_path = trimmed_path
+
            _update_task(task_id, progress=25)
            _update_task(task_id, message="正在合成唇形 (LatentSync)...", progress=30)

--- a/backend/app/services/storage.py
+++ b/backend/app/services/storage.py
@@ -20,12 +20,13 @@ class StorageService:
        self.BUCKET_MATERIALS = "materials"
        self.BUCKET_OUTPUTS = "outputs"
        self.BUCKET_REF_AUDIOS = "ref-audios"
+        self.BUCKET_GENERATED_AUDIOS = "generated-audios"
        # 确保所有 bucket 存在
        self._ensure_buckets()

    def _ensure_buckets(self):
        """确保所有必需的 bucket 存在"""
-        buckets = [self.BUCKET_MATERIALS, self.BUCKET_OUTPUTS, self.BUCKET_REF_AUDIOS]
+        buckets = [self.BUCKET_MATERIALS, self.BUCKET_OUTPUTS, self.BUCKET_REF_AUDIOS, self.BUCKET_GENERATED_AUDIOS]
        try:
            existing = self.supabase.storage.list_buckets()
            existing_names = {b.name for b in existing} if existing else set()
--- a/backend/app/services/video_service.py
+++ b/backend/app/services/video_service.py
@@ -210,9 +210,10 @@ class VideoService:
            return (0, 0)

    def prepare_segment(self, video_path: str, target_duration: float, output_path: str,
-                        target_resolution: tuple = None) -> str:
+                        target_resolution: tuple = None, source_start: float = 0.0) -> str:
        """将素材视频裁剪或循环到指定时长（无音频）。
        target_resolution: (width, height) 如需统一分辨率则传入，否则保持原分辨率。
+        source_start: 源视频截取起点（秒），默认 0。
        """
        Path(output_path).parent.mkdir(parents=True, exist_ok=True)

@@ -220,27 +221,62 @@ class VideoService:
        if video_dur <= 0:
            video_dur = target_duration

-        needs_loop = target_duration > video_dur
+        # 可用时长 = 从 source_start 到视频结尾
+        available = max(video_dur - source_start, 0.1)
+        needs_loop = target_duration > available
        needs_scale = target_resolution is not None

+        # 当需要循环且有 source_start 时，先裁剪出片段，再循环裁剪后的文件
+        # 避免 stream_loop 循环整个视频（而不是从 source_start 开始的片段）
+        actual_input = video_path
+        trim_temp = None
+        if needs_loop and source_start > 0:
+            trim_temp = str(Path(output_path).parent / (Path(output_path).stem + "_trim_tmp.mp4"))
+            trim_cmd = [
+                "ffmpeg", "-y",
+                "-ss", str(source_start),
+                "-i", video_path,
+                "-t", str(available),
+                "-an",
+                "-c:v", "libx264", "-preset", "fast", "-crf", "18",
+                trim_temp,
+            ]
+            if not self._run_ffmpeg(trim_cmd):
+                raise RuntimeError(f"FFmpeg trim for loop failed: {video_path}")
+            actual_input = trim_temp
+            source_start = 0.0  # 已裁剪，不需要再 seek
+            # 重新计算循环次数（基于裁剪后文件）
+            available = self._get_duration(trim_temp) or available
+
+        loop_count = int(target_duration / available) + 1 if needs_loop else 0
+
        cmd = ["ffmpeg", "-y"]
        if needs_loop:
-            loop_count = int(target_duration / video_dur) + 1
            cmd.extend(["-stream_loop", str(loop_count)])
-        cmd.extend(["-i", video_path, "-t", str(target_duration), "-an"])
+        if source_start > 0:
+            cmd.extend(["-ss", str(source_start)])
+        cmd.extend(["-i", actual_input, "-t", str(target_duration), "-an"])

        if needs_scale:
            w, h = target_resolution
            cmd.extend(["-vf", f"scale={w}:{h}:force_original_aspect_ratio=decrease,pad={w}:{h}:(ow-iw)/2:(oh-ih)/2"])

-        # 需要循环或缩放时必须重编码，否则用 stream copy 保持原画质
-        if needs_loop or needs_scale:
+        # 需要循环、缩放或指定起点时必须重编码，否则用 stream copy 保持原画质
+        if needs_loop or needs_scale or source_start > 0:
            cmd.extend(["-c:v", "libx264", "-preset", "fast", "-crf", "18"])
        else:
            cmd.extend(["-c:v", "copy"])

        cmd.append(output_path)

-        if self._run_ffmpeg(cmd):
-            return output_path
-        raise RuntimeError(f"FFmpeg prepare_segment failed: {video_path}")
+        try:
+            if self._run_ffmpeg(cmd):
+                return output_path
+            raise RuntimeError(f"FFmpeg prepare_segment failed: {video_path}")
+        finally:
+            # 清理裁剪临时文件
+            if trim_temp:
+                try:
+                    Path(trim_temp).unlink(missing_ok=True)
+                except Exception:
+                    pass
--- a/backend/scripts/watchdog.py
+++ b/backend/scripts/watchdog.py
@@ -23,20 +23,28 @@ SERVICES = [
        "name": "vigent2-qwen-tts",
        "url": "http://localhost:8009/health",
        "failures": 0,
-        "threshold": 3,
+        "threshold": 5,          # 连续5次失败才重启（5×30s = 2.5分钟容忍期）
        "timeout": 10.0,
-        "restart_cmd": ["pm2", "restart", "vigent2-qwen-tts"]
+        "restart_cmd": ["pm2", "restart", "vigent2-qwen-tts"],
+        "cooldown_until": 0,     # 重启后的冷却截止时间戳
+        "cooldown_sec": 120,     # 重启后等待120秒再开始检查
    }
 ]

 async def check_service(service):
    """检查单个服务健康状态"""
+    # 冷却期内跳过检查
+    now = time.time()
+    if now < service.get("cooldown_until", 0):
+        remaining = int(service["cooldown_until"] - now)
+        logger.debug(f"⏳ 服务 {service['name']} 冷却中，剩余 {remaining}s")
+        return True
+
    try:
        timeout = service.get("timeout", 10.0)
        async with httpx.AsyncClient(timeout=timeout) as client:
            response = await client.get(service["url"])
            if response.status_code == 200:
-                # 成功
                if service["failures"] > 0:
                    logger.info(f"✅ 服务 {service['name']} 已恢复正常")
                service["failures"] = 0
@@ -45,35 +53,36 @@ async def check_service(service):
                logger.warning(f"⚠️ 服务 {service['name']} 返回状态码 {response.status_code}")
    except Exception as e:
        logger.warning(f"⚠️ 无法连接服务 {service['name']}: {str(e)}")
-    
+
    # 失败处理
    service["failures"] += 1
    logger.warning(f"❌ 服务 {service['name']} 连续失败 {service['failures']}/{service['threshold']} 次")
-    
+
    if service["failures"] >= service['threshold']:
        logger.error(f"🚨 服务 {service['name']} 已达到失败阈值，正在重启...")
        try:
            subprocess.run(service["restart_cmd"], check=True)
            logger.info(f"♻️ 服务 {service['name']} 重启命令已发送")
-            # 重启后给予一段宽限期 (例如 60秒) 不检查，等待服务启动
-            service["failures"] = 0 # 重置计数
-            return "restarting" 
+            service["failures"] = 0
+            # 设置冷却期，等待服务完成启动和模型加载
+            service["cooldown_until"] = time.time() + service.get("cooldown_sec", 120)
+            return "restarting"
        except Exception as restart_error:
            logger.error(f"💥 重启服务 {service['name']} 失败: {restart_error}")
-            
+
    return False

 async def main():
    logger.info("🛡️ ViGent2 服务看门狗 (Watchdog) 已启动")
-    
+
+    # 启动时给所有服务一个初始冷却期，避免服务还没起来就被判定失败
+    for service in SERVICES:
+        service["cooldown_until"] = time.time() + 60
+
    while True:
-        # 并发检查所有服务
        for service in SERVICES:
-            result = await check_service(service)
-            if result == "restarting":
-                # 如果有服务重启，额外等待包含启动时间
-                pass
-                
+            await check_service(service)
+
        # 每 30 秒检查一次
        await asyncio.sleep(30)

--- a/frontend/package-lock.json
+++ b/frontend/package-lock.json
@@ -18,7 +18,8 @@
        "react": "19.2.3",
        "react-dom": "19.2.3",
        "sonner": "^2.0.7",
-        "swr": "^2.3.8"
+        "swr": "^2.3.8",
+        "wavesurfer.js": "^7.12.1"
      },
      "devDependencies": {
        "@tailwindcss/postcss": "^4",
@@ -6667,6 +6668,12 @@
        "react": "^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0"
      }
    },
+    "node_modules/wavesurfer.js": {
+      "version": "7.12.1",
+      "resolved": "https://registry.npmjs.org/wavesurfer.js/-/wavesurfer.js-7.12.1.tgz",
+      "integrity": "sha512-NswPjVHxk0Q1F/VMRemCPUzSojjuHHisQrBqQiRXg7MVbe3f5vQ6r0rTTXA/a/neC/4hnOEC4YpXca4LpH0SUg==",
+      "license": "BSD-3-Clause"
+    },
    "node_modules/which": {
      "version": "2.0.2",
      "resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz",
--- a/frontend/package.json
+++ b/frontend/package.json
@@ -19,7 +19,8 @@
    "react": "19.2.3",
    "react-dom": "19.2.3",
    "sonner": "^2.0.7",
-    "swr": "^2.3.8"
+    "swr": "^2.3.8",
+    "wavesurfer.js": "^7.12.1"
  },
  "devDependencies": {
    "@tailwindcss/postcss": "^4",
--- a/frontend/src/app/layout.tsx
+++ b/frontend/src/app/layout.tsx
@@ -46,7 +46,6 @@ export default function RootLayout({
        <Toaster
          position="top-center"
          richColors
-          closeButton
          toastOptions={{
            duration: 3000,
            className: "text-sm",
--- a/frontend/src/features/home/model/useGeneratedAudios.ts
+++ b/frontend/src/features/home/model/useGeneratedAudios.ts
@@ -0,0 +1,192 @@
+import { useCallback, useEffect, useRef, useState } from "react";
+import api from "@/shared/api/axios";
+import { ApiResponse, unwrap } from "@/shared/api/types";
+import { toast } from "sonner";
+
+export interface GeneratedAudio {
+  id: string;
+  name: string;
+  path: string;
+  duration_sec: number;
+  text: string;
+  tts_mode: string;
+  language: string;
+  created_at: number;
+}
+
+interface AudioTask {
+  status: string;
+  progress?: number;
+  message?: string;
+  output?: GeneratedAudio & { audio_id: string };
+}
+
+interface UseGeneratedAudiosOptions {
+  selectedAudioId: string | null;
+  setSelectedAudioId: React.Dispatch<React.SetStateAction<string | null>>;
+}
+
+export const useGeneratedAudios = ({
+  selectedAudioId,
+  setSelectedAudioId,
+}: UseGeneratedAudiosOptions) => {
+  const [generatedAudios, setGeneratedAudios] = useState<GeneratedAudio[]>([]);
+  const [selectedAudio, setSelectedAudio] = useState<GeneratedAudio | null>(null);
+  const [isGeneratingAudio, setIsGeneratingAudio] = useState(false);
+  const [audioTaskId, setAudioTaskId] = useState<string | null>(null);
+  const [audioTask, setAudioTask] = useState<AudioTask | null>(null);
+  const pollRef = useRef<NodeJS.Timeout | null>(null);
+
+  const fetchGeneratedAudios = useCallback(async (selectId?: string) => {
+    try {
+      const { data: res } = await api.get<ApiResponse<{ items: GeneratedAudio[] }>>(
+        "/api/generated-audios"
+      );
+      const payload = unwrap(res);
+      const items: GeneratedAudio[] = payload.items || [];
+      setGeneratedAudios(items);
+
+      if (selectId && items.length > 0) {
+        if (selectId === "__latest__") {
+          setSelectedAudioId(items[0].id);
+          setSelectedAudio(items[0]);
+        } else {
+          const found = items.find((a) => a.id === selectId);
+          if (found) {
+            setSelectedAudioId(found.id);
+            setSelectedAudio(found);
+          }
+        }
+      }
+    } catch (error) {
+      console.error("获取配音列表失败:", error);
+    }
+  }, [setSelectedAudioId]);
+
+  // Sync selectedAudio when selectedAudioId changes externally (e.g. from persistence)
+  useEffect(() => {
+    if (!selectedAudioId || generatedAudios.length === 0) return;
+    const found = generatedAudios.find((a) => a.id === selectedAudioId);
+    if (found) {
+      setSelectedAudio(found);
+    }
+  }, [selectedAudioId, generatedAudios]);
+
+  const stopPolling = useCallback(() => {
+    if (pollRef.current) {
+      clearInterval(pollRef.current);
+      pollRef.current = null;
+    }
+  }, []);
+
+  const startPolling = useCallback((taskId: string) => {
+    stopPolling();
+    pollRef.current = setInterval(async () => {
+      try {
+        const { data: res } = await api.get<ApiResponse<AudioTask>>(
+          `/api/generated-audios/tasks/${taskId}`
+        );
+        const task = unwrap(res);
+        setAudioTask(task);
+
+        if (task.status === "completed") {
+          stopPolling();
+          setIsGeneratingAudio(false);
+          setAudioTaskId(null);
+          // Refresh list and select the new audio
+          await fetchGeneratedAudios("__latest__");
+          toast.success(task.message || "配音生成完成");
+        } else if (task.status === "failed") {
+          stopPolling();
+          setIsGeneratingAudio(false);
+          setAudioTaskId(null);
+          toast.error(task.message || "配音生成失败");
+        } else if (task.status === "not_found") {
+          stopPolling();
+          setIsGeneratingAudio(false);
+          setAudioTaskId(null);
+          setAudioTask(null);
+          toast.error("任务已丢失（服务可能已重启），请重新生成");
+        }
+      } catch {
+        // Network error, keep polling
+      }
+    }, 1000);
+  }, [stopPolling, fetchGeneratedAudios]);
+
+  // Cleanup on unmount
+  useEffect(() => {
+    return () => stopPolling();
+  }, [stopPolling]);
+
+  const generateAudio = useCallback(async (params: {
+    text: string;
+    tts_mode: string;
+    voice?: string;
+    ref_audio_id?: string;
+    ref_text?: string;
+    language: string;
+  }) => {
+    setIsGeneratingAudio(true);
+    setAudioTask({ status: "pending", progress: 0, message: "正在提交..." });
+
+    try {
+      const { data: res } = await api.post<ApiResponse<{ task_id: string }>>(
+        "/api/generated-audios/generate",
+        params
+      );
+      const { task_id } = unwrap(res);
+      setAudioTaskId(task_id);
+      startPolling(task_id);
+    } catch (err: unknown) {
+      setIsGeneratingAudio(false);
+      setAudioTask(null);
+      const axiosErr = err as { response?: { data?: { message?: string } }; message?: string };
+      const errorMsg = axiosErr.response?.data?.message || axiosErr.message || String(err);
+      toast.error(`配音生成失败: ${errorMsg}`);
+    }
+  }, [startPolling]);
+
+  const deleteAudio = useCallback(async (audioId: string) => {
+    if (!confirm("确定要删除这个配音吗？")) return;
+    try {
+      await api.delete(`/api/generated-audios/${encodeURIComponent(audioId)}`);
+      if (selectedAudioId === audioId) {
+        setSelectedAudioId(null);
+        setSelectedAudio(null);
+      }
+      fetchGeneratedAudios();
+    } catch (error) {
+      toast.error("删除失败: " + error);
+    }
+  }, [fetchGeneratedAudios, selectedAudioId, setSelectedAudioId]);
+
+  const renameAudio = useCallback(async (audioId: string, newName: string) => {
+    try {
+      await api.put(`/api/generated-audios/${encodeURIComponent(audioId)}`, {
+        new_name: newName,
+      });
+      fetchGeneratedAudios();
+    } catch (err: unknown) {
+      toast.error("重命名失败: " + String(err));
+    }
+  }, [fetchGeneratedAudios]);
+
+  const selectAudio = useCallback((audio: GeneratedAudio) => {
+    setSelectedAudioId(audio.id);
+    setSelectedAudio(audio);
+  }, [setSelectedAudioId]);
+
+  return {
+    generatedAudios,
+    selectedAudio,
+    selectedAudioId,
+    isGeneratingAudio,
+    audioTask,
+    fetchGeneratedAudios,
+    generateAudio,
+    deleteAudio,
+    renameAudio,
+    selectAudio,
+  };
+};
--- a/frontend/src/features/home/model/useHomeController.ts
+++ b/frontend/src/features/home/model/useHomeController.ts
@@ -18,11 +18,14 @@ import { usePublishPrefetch } from "@/shared/hooks/usePublishPrefetch";
 import { PublishAccount } from "@/shared/types/publish";
 import { useBgm } from "@/features/home/model/useBgm";
 import { useGeneratedVideos } from "@/features/home/model/useGeneratedVideos";
+import { useGeneratedAudios } from "@/features/home/model/useGeneratedAudios";
 import { useHomePersistence } from "@/features/home/model/useHomePersistence";
 import { useMaterials } from "@/features/home/model/useMaterials";
 import { useMediaPlayers } from "@/features/home/model/useMediaPlayers";
 import { useRefAudios } from "@/features/home/model/useRefAudios";
 import { useTitleSubtitleStyles } from "@/features/home/model/useTitleSubtitleStyles";
+import { useTimelineEditor } from "@/features/home/model/useTimelineEditor";
+import { useSavedScripts } from "@/features/home/model/useSavedScripts";
 import { ApiResponse, unwrap } from "@/shared/api/types";

 const VOICES: Record<string, { id: string; name: string }[]> = {
@@ -164,6 +167,13 @@ export const useHomeController = () => {
  const [selectedRefAudio, setSelectedRefAudio] = useState<RefAudio | null>(null);
  const [refText, setRefText] = useState(FIXED_REF_TEXT);

+  // 预生成配音选中 ID
+  const [selectedAudioId, setSelectedAudioId] = useState<string | null>(null);
+
+  // ClipTrimmer 模态框状态
+  const [clipTrimmerOpen, setClipTrimmerOpen] = useState(false);
+  const [clipTrimmerSegmentId, setClipTrimmerSegmentId] = useState<string | null>(null);
+
  // 音频预览与重命名状态
  const [editingAudioId, setEditingAudioId] = useState<string | null>(null);
  const [editName, setEditName] = useState("");
@@ -347,6 +357,33 @@ export const useHomeController = () => {
    resolveMediaUrl,
  });

+  const {
+    generatedAudios,
+    selectedAudio,
+    isGeneratingAudio,
+    audioTask,
+    fetchGeneratedAudios,
+    generateAudio,
+    deleteAudio,
+    renameAudio,
+    selectAudio,
+  } = useGeneratedAudios({
+    selectedAudioId,
+    setSelectedAudioId,
+  });
+
+  const {
+    segments: timelineSegments,
+    reorderSegments,
+    setSourceRange,
+    toCustomAssignments,
+  } = useTimelineEditor({
+    audioDuration: selectedAudio?.duration_sec ?? 0,
+    materials,
+    selectedMaterials,
+    storageKey,
+  });
+
  useEffect(() => {
    if (isAuthLoading || !userId) return;
    let active = true;
@@ -420,8 +457,18 @@ export const useHomeController = () => {
    selectedVideoId,
    setSelectedVideoId,
    selectedRefAudio,
+    selectedAudioId,
+    setSelectedAudioId,
  });

+  const { savedScripts, saveScript, deleteScript: deleteSavedScript } = useSavedScripts(storageKey);
+
+  const handleSaveScript = () => {
+    if (!text.trim()) return;
+    saveScript(text);
+    toast.success("文案已保存");
+  };
+
  const syncTitleToPublish = (value: string) => {
    if (typeof window !== "undefined") {
      localStorage.setItem(`vigent_${storageKey}_publish_title`, value);
@@ -441,6 +488,7 @@ export const useHomeController = () => {
      fetchMaterials(),
      fetchGeneratedVideos(),
      fetchRefAudios(),
+      fetchGeneratedAudios(),
      refreshSubtitleStyles(),
      refreshTitleStyles(),
      fetchBgmList(),
@@ -537,14 +585,22 @@ export const useHomeController = () => {
    }
  }, [selectedBgmId, bgmList]);

+  // 素材列表滚动：跳过首次恢复，仅用户主动操作时滚动
+  const materialScrollReady = useRef(false);
  useEffect(() => {
    const firstSelected = selectedMaterials[0];
    if (!firstSelected) return;
+    if (!materialScrollReady.current) {
+      // 首次有选中素材时标记就绪，但不滚动（避免刷新后整页跳动）
+      materialScrollReady.current = true;
+      return;
+    }
    const target = materialItemRefs.current[firstSelected];
    if (target) {
      target.scrollIntoView({ block: "nearest", behavior: "smooth" });
    }
-  }, [selectedMaterials, materials]);
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [selectedMaterials.length]);

  // 【修复】历史视频默认选中逻辑
  // 当持久化恢复完成，且列表加载完毕，如果没选中任何视频，默认选中第一个
@@ -741,6 +797,28 @@ export const useHomeController = () => {
    }
  };

+  // 生成配音
+  const handleGenerateAudio = async () => {
+    if (!text.trim()) {
+      toast.error("请先输入文案");
+      return;
+    }
+    if (ttsMode === "voiceclone" && !selectedRefAudio) {
+      toast.error("请选择参考音频");
+      return;
+    }
+
+    const params = {
+      text: text.trim(),
+      tts_mode: ttsMode,
+      voice: ttsMode === "edgetts" ? voice : undefined,
+      ref_audio_id: ttsMode === "voiceclone" ? selectedRefAudio!.id : undefined,
+      ref_text: ttsMode === "voiceclone" ? refText : undefined,
+      language: textLang,
+    };
+    await generateAudio(params);
+  };
+
  // 生成视频
  const handleGenerate = async () => {
    if (selectedMaterials.length === 0 || !text.trim()) {
@@ -748,12 +826,9 @@ export const useHomeController = () => {
      return;
    }

-    // 声音克隆模式校验
-    if (ttsMode === "voiceclone") {
-      if (!selectedRefAudio) {
-        toast.error("请选择或上传参考音频");
-        return;
-      }
+    if (!selectedAudio) {
+      toast.error("请先生成并选中配音");
+      return;
    }

    if (enableBgm && !selectedBgmId) {
@@ -771,11 +846,12 @@ export const useHomeController = () => {
        return;
      }

-      // 构建请求参数
+      // 构建请求参数 - 使用预生成配音
      const payload: Record<string, unknown> = {
        material_path: firstMaterialObj.path,
-        text: text,
-        tts_mode: ttsMode,
+        text: selectedAudio.text || text,
+        generated_audio_id: selectedAudio.id,
+        language: selectedAudio.language || textLang,
        title: videoTitle.trim() || undefined,
        enable_subtitles: true,
      };
@@ -785,6 +861,16 @@ export const useHomeController = () => {
        payload.material_paths = selectedMaterials
          .map((id) => materials.find((x) => x.id === id)?.path)
          .filter((path): path is string => !!path);
+        // 发送自定义时间轴分配
+        const assignments = toCustomAssignments();
+        if (assignments.length > 0) {
+          payload.custom_assignments = assignments;
+        }
+      }
+
+      // 单素材 + 截取起点
+      if (selectedMaterials.length === 1 && timelineSegments[0]?.sourceStart > 0) {
+        payload.custom_assignments = toCustomAssignments();
      }

      if (selectedSubtitleStyleId) {
@@ -814,15 +900,6 @@ export const useHomeController = () => {
        payload.bgm_volume = bgmVolume;
      }

-      payload.language = textLang;
-
-      if (ttsMode === "edgetts") {
-        payload.voice = voice;
-      } else {
-        payload.ref_audio_id = selectedRefAudio!.id;
-        payload.ref_text = refText;
-      }
-
      // 创建生成任务
      const { data: res } = await api.post<ApiResponse<{ task_id: string }>>(
        "/api/videos/generate",
@@ -885,7 +962,6 @@ export const useHomeController = () => {
    handleUpload,
    selectedMaterials,
    toggleMaterial,
-    reorderMaterials,
    handlePreviewMaterial,
    editingMaterialId,
    editMaterialName,
@@ -903,6 +979,9 @@ export const useHomeController = () => {
    isTranslating,
    originalText,
    handleRestoreOriginal,
+    savedScripts,
+    handleSaveScript,
+    deleteSavedScript,
    showStylePreview,
    setShowStylePreview,
    videoTitle,
@@ -983,5 +1062,22 @@ export const useHomeController = () => {
    fetchGeneratedVideos,
    registerVideoRef,
    formatDate,
+    generatedAudios,
+    selectedAudio,
+    selectedAudioId,
+    isGeneratingAudio,
+    audioTask,
+    fetchGeneratedAudios,
+    handleGenerateAudio,
+    deleteAudio,
+    renameAudio,
+    selectAudio,
+    timelineSegments,
+    reorderSegments,
+    setSourceRange,
+    clipTrimmerOpen,
+    setClipTrimmerOpen,
+    clipTrimmerSegmentId,
+    setClipTrimmerSegmentId,
  };
 };
--- a/frontend/src/features/home/model/useHomePersistence.ts
+++ b/frontend/src/features/home/model/useHomePersistence.ts
@@ -48,6 +48,8 @@ interface UseHomePersistenceOptions {
  selectedVideoId: string | null;
  setSelectedVideoId: React.Dispatch<React.SetStateAction<string | null>>;
  selectedRefAudio: RefAudio | null;
+  selectedAudioId: string | null;
+  setSelectedAudioId: React.Dispatch<React.SetStateAction<string | null>>;
 }

 export const useHomePersistence = ({
@@ -88,6 +90,8 @@ export const useHomePersistence = ({
  selectedVideoId,
  setSelectedVideoId,
  selectedRefAudio,
+  selectedAudioId,
+  setSelectedAudioId,
 }: UseHomePersistenceOptions) => {
  const [isRestored, setIsRestored] = useState(false);

@@ -106,6 +110,7 @@ export const useHomePersistence = ({
    const savedTitleFontSize = localStorage.getItem(`vigent_${storageKey}_titleFontSize`);
    const savedBgmId = localStorage.getItem(`vigent_${storageKey}_bgmId`);
    const savedSelectedVideoId = localStorage.getItem(`vigent_${storageKey}_selectedVideoId`);
+    const savedSelectedAudioId = localStorage.getItem(`vigent_${storageKey}_selectedAudioId`);
    const savedBgmVolume = localStorage.getItem(`vigent_${storageKey}_bgmVolume`);
    const savedEnableBgm = localStorage.getItem(`vigent_${storageKey}_enableBgm`);
    const savedTitleTopMargin = localStorage.getItem(`vigent_${storageKey}_titleTopMargin`);
@@ -153,6 +158,7 @@ export const useHomePersistence = ({
    if (savedBgmVolume) setBgmVolume(parseFloat(savedBgmVolume));
    if (savedEnableBgm !== null) setEnableBgm(savedEnableBgm === 'true');
    if (savedSelectedVideoId) setSelectedVideoId(savedSelectedVideoId);
+    if (savedSelectedAudioId) setSelectedAudioId(savedSelectedAudioId);

    if (savedTitleTopMargin) {
      const parsed = parseInt(savedTitleTopMargin, 10);
@@ -174,6 +180,7 @@ export const useHomePersistence = ({
    setSelectedSubtitleStyleId,
    setSelectedTitleStyleId,
    setSelectedVideoId,
+    setSelectedAudioId,
    setSubtitleFontSize,
    setSubtitleSizeLocked,
    setText,
@@ -287,6 +294,15 @@ export const useHomePersistence = ({
    }
  }, [selectedVideoId, storageKey, isRestored]);

+  useEffect(() => {
+    if (!isRestored) return;
+    if (selectedAudioId) {
+      localStorage.setItem(`vigent_${storageKey}_selectedAudioId`, selectedAudioId);
+    } else {
+      localStorage.removeItem(`vigent_${storageKey}_selectedAudioId`);
+    }
+  }, [selectedAudioId, storageKey, isRestored]);
+
  useEffect(() => {
    if (isRestored && selectedRefAudio) {
      localStorage.setItem(`vigent_${storageKey}_refAudioId`, selectedRefAudio.id);
--- a/frontend/src/features/home/model/useMaterials.ts
+++ b/frontend/src/features/home/model/useMaterials.ts
@@ -2,8 +2,36 @@ import { useCallback, useState } from "react";
 import api from "@/shared/api/axios";
 import { ApiResponse, unwrap } from "@/shared/api/types";
 import { toast } from "sonner";
+import { resolveMediaUrl } from "@/shared/lib/media";
 import type { Material } from "@/shared/types/material";

+/** Probe video duration from a URL using <video> element */
+function probeVideoDuration(url: string): Promise<number> {
+  return new Promise((resolve) => {
+    const video = document.createElement("video");
+    video.preload = "metadata";
+    video.crossOrigin = "anonymous";
+    const cleanup = () => {
+      video.removeEventListener("loadedmetadata", onMeta);
+      video.removeEventListener("error", onError);
+      video.src = "";
+    };
+    const onMeta = () => {
+      const dur = video.duration;
+      cleanup();
+      resolve(Number.isFinite(dur) ? dur : 0);
+    };
+    const onError = () => {
+      cleanup();
+      resolve(0);
+    };
+    video.addEventListener("loadedmetadata", onMeta);
+    video.addEventListener("error", onError);
+    video.src = url;
+    video.load();
+  });
+}
+
 interface UseMaterialsOptions {
  selectedMaterials: string[];
  setSelectedMaterials: React.Dispatch<React.SetStateAction<string[]>>;
@@ -34,6 +62,18 @@ export const useMaterials = ({
      setMaterials(nextMaterials);
      setLastMaterialCount(nextMaterials.length);

+      // Probe video durations in background
+      if (nextMaterials.length > 0) {
+        Promise.all(
+          nextMaterials.map(async (m) => {
+            const url = resolveMediaUrl(m.path);
+            if (!url) return m;
+            const dur = await probeVideoDuration(url);
+            return { ...m, duration_sec: dur };
+          })
+        ).then((enriched) => setMaterials(enriched));
+      }
+
      setSelectedMaterials((prev) => {
        // 保留已选中且仍存在的
        const existingIds = new Set(nextMaterials.map((m) => m.id));
@@ -133,6 +173,18 @@ export const useMaterials = ({
      setMaterials(nextMaterials);
      setLastMaterialCount(nextMaterials.length);

+      // Probe video durations in background
+      if (nextMaterials.length > 0) {
+        Promise.all(
+          nextMaterials.map(async (m) => {
+            const url = resolveMediaUrl(m.path);
+            if (!url) return m;
+            const dur = await probeVideoDuration(url);
+            return { ...m, duration_sec: dur };
+          })
+        ).then((enriched) => setMaterials(enriched));
+      }
+
      // 找出新增的素材 ID 并自动选中
      const oldIds = new Set(materials.map((m) => m.id));
      const newIds = nextMaterials.filter((m) => !oldIds.has(m.id)).map((m) => m.id);
--- a/frontend/src/features/home/model/useSavedScripts.ts
+++ b/frontend/src/features/home/model/useSavedScripts.ts
@@ -0,0 +1,51 @@
+import { useState, useEffect, useRef } from "react";
+
+export interface SavedScript {
+  id: string;
+  name: string;
+  content: string;
+  savedAt: number;
+}
+
+export function useSavedScripts(storageKey: string) {
+  const lsKey = `vigent_${storageKey}_savedScripts`;
+  const lsKeyRef = useRef(lsKey);
+  lsKeyRef.current = lsKey;
+
+  const [savedScripts, setSavedScripts] = useState<SavedScript[]>([]);
+
+  // Re-read from localStorage whenever lsKey changes (e.g. guest → userId)
+  useEffect(() => {
+    try {
+      const raw = localStorage.getItem(lsKey);
+      setSavedScripts(raw ? JSON.parse(raw) : []);
+    } catch {
+      setSavedScripts([]);
+    }
+  }, [lsKey]);
+
+  const saveScript = (content: string) => {
+    const name = content.slice(0, 15).replace(/\n/g, " ") || "未命名";
+    const entry: SavedScript = {
+      id: Date.now().toString(36) + Math.random().toString(36).slice(2, 6),
+      name,
+      content,
+      savedAt: Date.now(),
+    };
+    setSavedScripts((prev) => {
+      const next = [entry, ...prev];
+      localStorage.setItem(lsKeyRef.current, JSON.stringify(next));
+      return next;
+    });
+  };
+
+  const deleteScript = (id: string) => {
+    setSavedScripts((prev) => {
+      const next = prev.filter((s) => s.id !== id);
+      localStorage.setItem(lsKeyRef.current, JSON.stringify(next));
+      return next;
+    });
+  };
+
+  return { savedScripts, saveScript, deleteScript };
+}
--- a/frontend/src/features/home/model/useTimelineEditor.ts
+++ b/frontend/src/features/home/model/useTimelineEditor.ts
@@ -0,0 +1,246 @@
+import { useCallback, useEffect, useRef, useState } from "react";
+import type { Material } from "@/shared/types/material";
+
+export interface TimelineSegment {
+  id: string;
+  materialId: string;
+  materialName: string;
+  start: number;
+  end: number;
+  sourceStart: number;
+  sourceEnd: number;
+  color: string;
+}
+
+export interface CustomAssignment {
+  material_path: string;
+  start: number;
+  end: number;
+  source_start: number;
+}
+
+const COLORS = ["#8b5cf6", "#ec4899", "#06b6d4", "#f59e0b", "#10b981", "#f97316"];
+
+/** Serializable subset for localStorage */
+interface SegmentSnapshot {
+  materialId: string;
+  start: number;
+  end: number;
+  sourceStart: number;
+  sourceEnd: number;
+}
+
+/** Get effective duration of a segment (clipped range or full material duration) */
+function getEffectiveDuration(
+  seg: { sourceStart: number; sourceEnd: number; materialId: string },
+  mats: Material[]
+): number {
+  if (seg.sourceEnd > seg.sourceStart) return seg.sourceEnd - seg.sourceStart;
+  const mat = mats.find((m) => m.id === seg.materialId);
+  return mat?.duration_sec ?? 0;
+}
+
+/**
+ * Recalculate segment start/end positions based on effective durations.
+ * - Segments placed sequentially by effective duration
+ * - Segments exceeding audioDuration keep their positions (overflow, start >= duration)
+ * - Last visible segment is capped/extended to exactly audioDuration (loop fill)
+ */
+function recalcPositions(
+  segs: TimelineSegment[],
+  mats: Material[],
+  duration: number
+): TimelineSegment[] {
+  if (segs.length === 0 || duration <= 0) return segs;
+
+  const fallbackDur = duration / segs.length;
+  let cursor = 0;
+  const result = segs.map((seg) => {
+    const effDur = getEffectiveDuration(seg, mats);
+    const dur = effDur > 0 ? effDur : fallbackDur;
+    const newSeg = { ...seg, start: cursor, end: cursor + dur };
+    cursor += dur;
+    return newSeg;
+  });
+
+  // Find last segment that starts before audioDuration
+  let lastVisibleIdx = -1;
+  for (let i = result.length - 1; i >= 0; i--) {
+    if (result[i].start < duration) {
+      lastVisibleIdx = i;
+      break;
+    }
+  }
+
+  // Cap/extend last visible segment to exactly audioDuration
+  if (lastVisibleIdx >= 0) {
+    result[lastVisibleIdx] = { ...result[lastVisibleIdx], end: duration };
+  }
+
+  return result;
+}
+
+interface UseTimelineEditorOptions {
+  audioDuration: number;
+  materials: Material[];
+  selectedMaterials: string[];
+  storageKey?: string;
+}
+
+export const useTimelineEditor = ({
+  audioDuration,
+  materials,
+  selectedMaterials,
+  storageKey,
+}: UseTimelineEditorOptions) => {
+  const [segments, setSegments] = useState<TimelineSegment[]>([]);
+  const prevKey = useRef("");
+  const restoredRef = useRef(false);
+
+  // Refs for stable callbacks (avoid recreating on every materials/duration change)
+  const materialsRef = useRef(materials);
+  materialsRef.current = materials;
+  const audioDurationRef = useRef(audioDuration);
+  audioDurationRef.current = audioDuration;
+
+  // Build a durationsKey so segments re-init when material durations become available
+  const durationsKey = selectedMaterials
+    .map((id) => materials.find((m) => m.id === id)?.duration_sec ?? 0)
+    .join(",");
+
+  // Build a cache key from materials + duration
+  const cacheKey = `${selectedMaterials.join(",")}_${audioDuration.toFixed(1)}`;
+  const lsKey = storageKey ? `vigent_${storageKey}_timeline` : null;
+
+  const initSegments = useCallback(() => {
+    if (selectedMaterials.length === 0 || audioDuration <= 0) {
+      setSegments([]);
+      return;
+    }
+
+    // Try restore from localStorage
+    if (lsKey) {
+      try {
+        const raw = localStorage.getItem(lsKey);
+        if (raw) {
+          const saved = JSON.parse(raw) as { key: string; segments: SegmentSnapshot[] };
+          if (saved.key === cacheKey && saved.segments.length === selectedMaterials.length) {
+            const allMatch = saved.segments.every(
+              (s, i) => s.materialId === selectedMaterials[i] || saved.segments.some((ss) => ss.materialId === selectedMaterials[i])
+            );
+            if (allMatch) {
+              const restored: TimelineSegment[] = saved.segments.map((s, i) => {
+                const mat = materials.find((m) => m.id === s.materialId);
+                return {
+                  id: `seg-${i}-${Date.now()}`,
+                  materialId: s.materialId,
+                  materialName: mat?.scene || mat?.name || s.materialId,
+                  start: 0,
+                  end: 0,
+                  sourceStart: s.sourceStart,
+                  sourceEnd: s.sourceEnd,
+                  color: COLORS[i % COLORS.length],
+                };
+              });
+              setSegments(recalcPositions(restored, materials, audioDuration));
+              restoredRef.current = true;
+              return;
+            }
+          }
+        }
+      } catch {
+        // ignore parse errors
+      }
+    }
+
+    // Create fresh segments — positions derived by recalcPositions
+    const newSegments: TimelineSegment[] = selectedMaterials.map((matId, i) => {
+      const mat = materials.find((m) => m.id === matId);
+      return {
+        id: `seg-${i}-${Date.now()}`,
+        materialId: matId,
+        materialName: mat?.scene || mat?.name || matId,
+        start: 0,
+        end: 0,
+        sourceStart: 0,
+        sourceEnd: 0,
+        color: COLORS[i % COLORS.length],
+      };
+    });
+
+    setSegments(recalcPositions(newSegments, materials, audioDuration));
+  }, [audioDuration, materials, selectedMaterials, lsKey, cacheKey]);
+
+  // Auto-init when selectedMaterials, audioDuration, or material durations change
+  useEffect(() => {
+    const key = `${selectedMaterials.join(",")}_${audioDuration}_${durationsKey}`;
+    if (key !== prevKey.current) {
+      prevKey.current = key;
+      initSegments();
+    }
+  }, [selectedMaterials, audioDuration, durationsKey, initSegments]);
+
+  // Persist segments to localStorage on change (debounced)
+  useEffect(() => {
+    if (!lsKey || segments.length === 0) return;
+    const timeout = setTimeout(() => {
+      const snapshots: SegmentSnapshot[] = segments.map((s) => ({
+        materialId: s.materialId,
+        start: s.start,
+        end: s.end,
+        sourceStart: s.sourceStart,
+        sourceEnd: s.sourceEnd,
+      }));
+      localStorage.setItem(lsKey, JSON.stringify({ key: cacheKey, segments: snapshots }));
+    }, 300);
+    return () => clearTimeout(timeout);
+  }, [segments, lsKey, cacheKey]);
+
+  const reorderSegments = useCallback(
+    (fromIdx: number, toIdx: number) => {
+      setSegments((prev) => {
+        if (fromIdx < 0 || toIdx < 0 || fromIdx >= prev.length || toIdx >= prev.length) return prev;
+        if (fromIdx === toIdx) return prev;
+        const next = [...prev];
+        // Move the segment: remove from old position, insert at new position
+        const [moved] = next.splice(fromIdx, 1);
+        next.splice(toIdx, 0, moved);
+        return recalcPositions(next, materialsRef.current, audioDurationRef.current);
+      });
+    },
+    []
+  );
+
+  const setSourceRange = useCallback(
+    (id: string, sourceStart: number, sourceEnd: number) => {
+      setSegments((prev) => {
+        const updated = prev.map((s) => (s.id === id ? { ...s, sourceStart, sourceEnd } : s));
+        return recalcPositions(updated, materialsRef.current, audioDurationRef.current);
+      });
+    },
+    []
+  );
+
+  const toCustomAssignments = useCallback((): CustomAssignment[] => {
+    const duration = audioDurationRef.current;
+    return segments
+      .filter((seg) => seg.start < duration)
+      .map((seg) => {
+        const mat = materialsRef.current.find((m) => m.id === seg.materialId);
+        return {
+          material_path: mat?.path || seg.materialId,
+          start: seg.start,
+          end: seg.end,
+          source_start: seg.sourceStart,
+        };
+      });
+  }, [segments]);
+
+  return {
+    segments,
+    initSegments,
+    reorderSegments,
+    setSourceRange,
+    toCustomAssignments,
+  };
+};
--- a/frontend/src/features/home/ui/ClipTrimmer.tsx
+++ b/frontend/src/features/home/ui/ClipTrimmer.tsx
@@ -0,0 +1,293 @@
+import { useCallback, useEffect, useRef, useState } from "react";
+import { X, Play, Pause } from "lucide-react";
+import type { TimelineSegment } from "@/features/home/model/useTimelineEditor";
+
+interface ClipTrimmerProps {
+  isOpen: boolean;
+  segment: TimelineSegment | null;
+  materialUrl: string | null;
+  onConfirm: (sourceStart: number, sourceEnd: number) => void;
+  onClose: () => void;
+}
+
+function formatSec(sec: number): string {
+  const m = Math.floor(sec / 60);
+  const s = sec % 60;
+  return `${String(m).padStart(2, "0")}:${s.toFixed(1).padStart(4, "0")}`;
+}
+
+export function ClipTrimmer({
+  isOpen,
+  segment,
+  materialUrl,
+  onConfirm,
+  onClose,
+}: ClipTrimmerProps) {
+  const videoRef = useRef<HTMLVideoElement>(null);
+  const trackRef = useRef<HTMLDivElement>(null);
+  const [duration, setDuration] = useState(0);
+  const [sourceStart, setSourceStart] = useState(0);
+  const [sourceEnd, setSourceEnd] = useState(0);
+  const [currentTime, setCurrentTime] = useState(0);
+  const [isPlaying, setIsPlaying] = useState(false);
+  const [dragging, setDragging] = useState<"start" | "end" | null>(null);
+  const animRef = useRef<number>(0);
+
+  // Reset state when segment changes
+  useEffect(() => {
+    if (segment && isOpen) {
+      setSourceStart(segment.sourceStart);
+      setSourceEnd(segment.sourceEnd);
+      setCurrentTime(segment.sourceStart);
+      setIsPlaying(false);
+    }
+  }, [segment, isOpen]);
+
+  // Track currentTime during playback
+  useEffect(() => {
+    if (!isPlaying || !videoRef.current) return;
+
+    const tick = () => {
+      if (!videoRef.current) return;
+      const t = videoRef.current.currentTime;
+      const end = sourceEnd || duration;
+      if (t >= end) {
+        videoRef.current.pause();
+        videoRef.current.currentTime = sourceStart;
+        setCurrentTime(sourceStart);
+        setIsPlaying(false);
+        return;
+      }
+      setCurrentTime(t);
+      animRef.current = requestAnimationFrame(tick);
+    };
+    animRef.current = requestAnimationFrame(tick);
+    return () => cancelAnimationFrame(animRef.current);
+  }, [isPlaying, sourceStart, sourceEnd, duration]);
+
+  // Seek video when not playing and currentTime changes
+  useEffect(() => {
+    if (videoRef.current && !isPlaying) {
+      videoRef.current.currentTime = currentTime;
+    }
+  }, [currentTime, isPlaying]);
+
+  const handleLoadedMetadata = useCallback(() => {
+    if (videoRef.current) {
+      const dur = videoRef.current.duration;
+      setDuration(dur);
+      if (sourceEnd === 0) {
+        setSourceEnd(dur);
+      }
+    }
+  }, [sourceEnd]);
+
+  const togglePlay = useCallback(() => {
+    if (!videoRef.current || duration === 0) return;
+    if (isPlaying) {
+      videoRef.current.pause();
+      setIsPlaying(false);
+    } else {
+      const end = sourceEnd || duration;
+      if (videoRef.current.currentTime >= end || videoRef.current.currentTime < sourceStart) {
+        videoRef.current.currentTime = sourceStart;
+        setCurrentTime(sourceStart);
+      }
+      videoRef.current.play().catch(() => {});
+      setIsPlaying(true);
+    }
+  }, [isPlaying, sourceStart, sourceEnd, duration]);
+
+  // --- Dual-handle slider logic ---
+  const getPositionFromEvent = useCallback(
+    (clientX: number) => {
+      if (!trackRef.current || duration === 0) return 0;
+      const rect = trackRef.current.getBoundingClientRect();
+      const ratio = Math.max(0, Math.min(1, (clientX - rect.left) / rect.width));
+      return ratio * duration;
+    },
+    [duration]
+  );
+
+  const handleThumbPointerDown = useCallback(
+    (which: "start" | "end", e: React.PointerEvent) => {
+      e.preventDefault();
+      e.stopPropagation();
+      setDragging(which);
+      (e.target as HTMLElement).setPointerCapture(e.pointerId);
+    },
+    []
+  );
+
+  const handleTrackPointerMove = useCallback(
+    (e: React.PointerEvent) => {
+      if (!dragging) return;
+      const pos = getPositionFromEvent(e.clientX);
+      const minGap = 0.5;
+      if (dragging === "start") {
+        const clamped = Math.max(0, Math.min(pos, (sourceEnd || duration) - minGap));
+        setSourceStart(clamped);
+        setCurrentTime(clamped);
+      } else {
+        const clamped = Math.min(duration, Math.max(pos, sourceStart + minGap));
+        setSourceEnd(clamped);
+      }
+    },
+    [dragging, getPositionFromEvent, sourceStart, sourceEnd, duration]
+  );
+
+  const handleTrackPointerUp = useCallback(() => {
+    setDragging(null);
+  }, []);
+
+  const handleConfirm = () => {
+    onConfirm(sourceStart, sourceEnd >= duration ? 0 : sourceEnd);
+  };
+
+  if (!isOpen || !segment) return null;
+
+  const assignedDur = segment.end - segment.start;
+  const effectiveEnd = sourceEnd || duration;
+  const clipDur = effectiveEnd - sourceStart;
+  const startPct = duration > 0 ? (sourceStart / duration) * 100 : 0;
+  const endPct = duration > 0 ? (effectiveEnd / duration) * 100 : 100;
+  const playheadPct = duration > 0 ? (currentTime / duration) * 100 : 0;
+
+  return (
+    <div className="fixed inset-0 z-50 flex items-center justify-center bg-black/60 backdrop-blur-sm" onClick={onClose}>
+      <div
+        className="bg-gray-900 border border-white/10 rounded-2xl w-full max-w-lg mx-4 overflow-hidden"
+        onClick={(e) => e.stopPropagation()}
+      >
+        {/* Header */}
+        <div className="flex items-center justify-between px-5 py-3 border-b border-white/10">
+          <h3 className="text-white font-semibold text-sm">
+            截取设置 - {segment.materialName}
+          </h3>
+          <button onClick={onClose} className="text-gray-400 hover:text-white">
+            <X className="h-4 w-4" />
+          </button>
+        </div>
+
+        {/* Video preview */}
+        <div className="px-5 pt-4">
+          <div className="relative bg-black rounded-lg overflow-hidden aspect-video group">
+            {materialUrl ? (
+              <video
+                ref={videoRef}
+                src={materialUrl}
+                className="w-full h-full object-contain"
+                onLoadedMetadata={handleLoadedMetadata}
+                onEnded={() => setIsPlaying(false)}
+                preload="auto"
+                muted
+              />
+            ) : (
+              <div className="flex items-center justify-center h-full text-gray-500 text-sm">
+                无法加载视频
+              </div>
+            )}
+            {/* Play/Pause overlay */}
+            {materialUrl && (
+              <button
+                onClick={togglePlay}
+                className="absolute inset-0 flex items-center justify-center bg-black/0 hover:bg-black/30 transition-colors"
+              >
+                <div className={`p-3 rounded-full bg-black/60 text-white transition-opacity ${isPlaying ? "opacity-0 group-hover:opacity-100" : "opacity-100"}`}>
+                  {isPlaying ? <Pause className="h-6 w-6" /> : <Play className="h-6 w-6" />}
+                </div>
+              </button>
+            )}
+            <div className="absolute bottom-2 right-2 bg-black/70 text-white text-[10px] px-2 py-0.5 rounded pointer-events-none">
+              {formatSec(currentTime)}
+            </div>
+          </div>
+        </div>
+
+        {/* Dual-handle range slider */}
+        <div className="px-5 py-4 space-y-3">
+          <div className="text-xs text-gray-400 flex justify-between">
+            <span>源视频时长: {duration > 0 ? formatSec(duration) : "加载中..."}</span>
+          </div>
+
+          {/* Custom range track */}
+          <div
+            ref={trackRef}
+            className="relative h-8 cursor-pointer select-none touch-none"
+            onPointerMove={handleTrackPointerMove}
+            onPointerUp={handleTrackPointerUp}
+            onPointerLeave={handleTrackPointerUp}
+          >
+            {/* Background track */}
+            <div className="absolute top-1/2 -translate-y-1/2 left-0 right-0 h-2 bg-white/10 rounded-full" />
+
+            {/* Selected range */}
+            <div
+              className="absolute top-1/2 -translate-y-1/2 h-2 rounded-full"
+              style={{
+                left: `${startPct}%`,
+                width: `${endPct - startPct}%`,
+                backgroundColor: segment.color + "88",
+              }}
+            />
+
+            {/* Playhead indicator */}
+            {duration > 0 && (
+              <div
+                className="absolute top-1/2 -translate-y-1/2 w-0.5 h-4 bg-white/60 rounded-full pointer-events-none"
+                style={{ left: `${playheadPct}%` }}
+              />
+            )}
+
+            {/* Start thumb */}
+            <div
+              onPointerDown={(e) => handleThumbPointerDown("start", e)}
+              className="absolute top-1/2 -translate-y-1/2 -translate-x-1/2 w-4 h-4 rounded-full bg-purple-500 border-2 border-white shadow-lg cursor-grab active:cursor-grabbing hover:scale-110 transition-transform z-10"
+              style={{ left: `${startPct}%` }}
+              title={`起点: ${formatSec(sourceStart)}`}
+            />
+
+            {/* End thumb */}
+            <div
+              onPointerDown={(e) => handleThumbPointerDown("end", e)}
+              className="absolute top-1/2 -translate-y-1/2 -translate-x-1/2 w-4 h-4 rounded-full bg-pink-500 border-2 border-white shadow-lg cursor-grab active:cursor-grabbing hover:scale-110 transition-transform z-10"
+              style={{ left: `${endPct}%` }}
+              title={`终点: ${formatSec(effectiveEnd)}`}
+            />
+          </div>
+
+          {/* Time labels */}
+          <div className="flex justify-between text-xs text-gray-400">
+            <span className="text-purple-400">{formatSec(sourceStart)}</span>
+            <span className="text-pink-400">{formatSec(effectiveEnd)}</span>
+          </div>
+
+          {/* Info */}
+          <div className="text-[11px] text-gray-500 flex items-center gap-2 flex-wrap">
+            <span>截取: {clipDur.toFixed(1)}s</span>
+            <span className="text-gray-600">|</span>
+            <span>分配: {assignedDur.toFixed(1)}s</span>
+            {clipDur < assignedDur && <span className="text-amber-500">(将循环补足)</span>}
+            {clipDur > assignedDur && <span className="text-cyan-500">(将截断)</span>}
+          </div>
+        </div>
+
+        {/* Actions */}
+        <div className="flex justify-end gap-2 px-5 pb-4">
+          <button
+            onClick={onClose}
+            className="px-4 py-1.5 text-xs bg-white/10 hover:bg-white/20 rounded-lg text-gray-300 transition-colors"
+          >
+            取消
+          </button>
+          <button
+            onClick={handleConfirm}
+            className="px-4 py-1.5 text-xs bg-gradient-to-r from-purple-600 to-pink-600 hover:from-purple-700 hover:to-pink-700 text-white rounded-lg transition-colors"
+          >
+            确定
+          </button>
+        </div>
+      </div>
+    </div>
+  );
+}
--- a/frontend/src/features/home/ui/GeneratedAudiosPanel.tsx
+++ b/frontend/src/features/home/ui/GeneratedAudiosPanel.tsx
@@ -0,0 +1,224 @@
+import { useState, useRef, useCallback, useEffect } from "react";
+import { Play, Pause, Pencil, Trash2, Check, X, RefreshCw, Mic } from "lucide-react";
+import type { GeneratedAudio } from "@/features/home/model/useGeneratedAudios";
+
+interface AudioTask {
+  status: string;
+  progress?: number;
+  message?: string;
+}
+
+interface GeneratedAudiosPanelProps {
+  generatedAudios: GeneratedAudio[];
+  selectedAudioId: string | null;
+  isGeneratingAudio: boolean;
+  audioTask: AudioTask | null;
+  onGenerateAudio: () => void;
+  onRefresh: () => void;
+  onSelectAudio: (audio: GeneratedAudio) => void;
+  onDeleteAudio: (id: string) => void;
+  onRenameAudio: (id: string, newName: string) => void;
+  hasText: boolean;
+}
+
+export function GeneratedAudiosPanel({
+  generatedAudios,
+  selectedAudioId,
+  isGeneratingAudio,
+  audioTask,
+  onGenerateAudio,
+  onRefresh,
+  onSelectAudio,
+  onDeleteAudio,
+  onRenameAudio,
+  hasText,
+}: GeneratedAudiosPanelProps) {
+  const [editingId, setEditingId] = useState<string | null>(null);
+  const [editName, setEditName] = useState("");
+  const [playingId, setPlayingId] = useState<string | null>(null);
+  const audioRef = useRef<HTMLAudioElement | null>(null);
+
+  const stopPlaying = useCallback(() => {
+    if (audioRef.current) {
+      audioRef.current.pause();
+      audioRef.current.currentTime = 0;
+      audioRef.current = null;
+    }
+    setPlayingId(null);
+  }, []);
+
+  // Cleanup on unmount
+  useEffect(() => {
+    return () => {
+      if (audioRef.current) {
+        audioRef.current.pause();
+        audioRef.current = null;
+      }
+    };
+  }, []);
+
+  const togglePlay = (audio: GeneratedAudio, e: React.MouseEvent) => {
+    e.stopPropagation();
+    if (playingId === audio.id) {
+      stopPlaying();
+      return;
+    }
+    stopPlaying();
+    const player = new Audio(audio.path);
+    player.onended = () => setPlayingId(null);
+    player.play().catch(() => {});
+    audioRef.current = player;
+    setPlayingId(audio.id);
+  };
+
+  const startEditing = (audio: GeneratedAudio, e: React.MouseEvent) => {
+    e.stopPropagation();
+    setEditingId(audio.id);
+    setEditName(audio.name);
+  };
+
+  const saveEditing = (audioId: string, e: React.MouseEvent) => {
+    e.stopPropagation();
+    if (!editName.trim()) return;
+    onRenameAudio(audioId, editName.trim());
+    setEditingId(null);
+    setEditName("");
+  };
+
+  const cancelEditing = (e: React.MouseEvent) => {
+    e.stopPropagation();
+    setEditingId(null);
+    setEditName("");
+  };
+
+  return (
+    <div className="bg-white/5 rounded-2xl p-4 sm:p-6 border border-white/10 backdrop-blur-sm">
+      <div className="flex justify-between items-center gap-2 mb-4">
+        <h2 className="text-base sm:text-lg font-semibold text-white flex items-center gap-2 whitespace-nowrap">
+          <Mic className="h-4 w-4 text-purple-400" />
+          配音列表
+        </h2>
+        <div className="flex gap-1.5">
+          <button
+            onClick={onGenerateAudio}
+            disabled={isGeneratingAudio || !hasText}
+            className={`px-2 py-1 text-xs rounded transition-all whitespace-nowrap flex items-center gap-1 ${
+              isGeneratingAudio || !hasText
+                ? "bg-gray-600 cursor-not-allowed text-gray-400"
+                : "bg-gradient-to-r from-purple-600 to-pink-600 hover:from-purple-700 hover:to-pink-700 text-white"
+            }`}
+          >
+            <Mic className="h-3.5 w-3.5" />
+            生成配音
+          </button>
+          <button
+            onClick={onRefresh}
+            className="px-2 py-1 text-xs bg-white/10 hover:bg-white/20 rounded text-gray-300 whitespace-nowrap flex items-center gap-1"
+          >
+            <RefreshCw className="h-3.5 w-3.5" />
+          </button>
+        </div>
+      </div>
+
+      {/* 生成进度 */}
+      {isGeneratingAudio && audioTask && (
+        <div className="mb-4 p-3 bg-purple-500/10 rounded-xl border border-purple-500/30">
+          <div className="flex justify-between text-sm text-purple-300 mb-2">
+            <span>{audioTask.message || "生成中..."}</span>
+            <span>{audioTask.progress || 0}%</span>
+          </div>
+          <div className="h-2 bg-black/30 rounded-full overflow-hidden">
+            <div
+              className="h-full bg-gradient-to-r from-purple-500 to-pink-500 transition-all duration-300"
+              style={{ width: `${audioTask.progress || 0}%` }}
+            />
+          </div>
+        </div>
+      )}
+
+      {/* 配音列表 */}
+      {generatedAudios.length === 0 ? (
+        <div className="text-center py-6 text-gray-400">
+          <p className="text-sm">暂无配音</p>
+          <p className="text-xs mt-1 text-gray-500">点击「生成配音」创建</p>
+        </div>
+      ) : (
+        <div className="space-y-2 max-h-48 sm:max-h-56 overflow-y-auto hide-scrollbar">
+          {generatedAudios.map((audio) => {
+            const isSelected = selectedAudioId === audio.id;
+            return (
+              <div
+                key={audio.id}
+                onClick={() => onSelectAudio(audio)}
+                className={`p-3 rounded-lg border transition-all cursor-pointer flex items-center justify-between group ${
+                  isSelected
+                    ? "border-purple-500 bg-purple-500/20"
+                    : "border-white/10 bg-white/5 hover:border-white/30"
+                }`}
+              >
+                {editingId === audio.id ? (
+                  <div className="flex-1 flex items-center gap-2" onClick={(e) => e.stopPropagation()}>
+                    <input
+                      value={editName}
+                      onChange={(e) => setEditName(e.target.value)}
+                      className="flex-1 bg-black/40 border border-white/20 rounded-md px-2 py-1 text-xs text-white"
+                      autoFocus
+                      onKeyDown={(e) => {
+                        if (e.key === "Enter") saveEditing(audio.id, e as unknown as React.MouseEvent);
+                        if (e.key === "Escape") cancelEditing(e as unknown as React.MouseEvent);
+                      }}
+                    />
+                    <button onClick={(e) => saveEditing(audio.id, e)} className="p-1 text-green-400 hover:text-green-300" title="保存">
+                      <Check className="h-4 w-4" />
+                    </button>
+                    <button onClick={cancelEditing} className="p-1 text-gray-400 hover:text-white" title="取消">
+                      <X className="h-4 w-4" />
+                    </button>
+                  </div>
+                ) : (
+                  <>
+                    <div className="min-w-0 flex-1">
+                      <div className="text-white text-sm truncate">{audio.name}</div>
+                      <div className="text-gray-400 text-xs">{audio.duration_sec.toFixed(1)}s</div>
+                    </div>
+                    <div className="flex items-center gap-1 pl-2 opacity-0 group-hover:opacity-100 transition-opacity">
+                      <button
+                        onClick={(e) => togglePlay(audio, e)}
+                        className="p-1 text-gray-500 hover:text-purple-400 transition-colors"
+                        title={playingId === audio.id ? "暂停" : "播放"}
+                      >
+                        {playingId === audio.id ? (
+                          <Pause className="h-3.5 w-3.5" />
+                        ) : (
+                          <Play className="h-3.5 w-3.5" />
+                        )}
+                      </button>
+                      <button
+                        onClick={(e) => startEditing(audio, e)}
+                        className="p-1 text-gray-500 hover:text-white transition-colors"
+                        title="重命名"
+                      >
+                        <Pencil className="h-3.5 w-3.5" />
+                      </button>
+                      <button
+                        onClick={(e) => {
+                          e.stopPropagation();
+                          onDeleteAudio(audio.id);
+                        }}
+                        className="p-1 text-gray-500 hover:text-red-400 transition-colors"
+                        title="删除"
+                      >
+                        <Trash2 className="h-3.5 w-3.5" />
+                      </button>
+                    </div>
+                  </>
+                )}
+              </div>
+            );
+          })}
+        </div>
+      )}
+
+    </div>
+  );
+}
--- a/frontend/src/features/home/ui/HomePage.tsx
+++ b/frontend/src/features/home/ui/HomePage.tsx
@@ -1,20 +1,24 @@
 "use client";

-import { useEffect } from "react";
+import { useEffect, useMemo } from "react";
 import { useRouter } from "next/navigation";
 import VideoPreviewModal from "@/components/VideoPreviewModal";
 import ScriptExtractionModal from "./ScriptExtractionModal";
 import { useHomeController } from "@/features/home/model/useHomeController";
+import { resolveMediaUrl } from "@/shared/lib/media";
 import { BgmPanel } from "@/features/home/ui/BgmPanel";
 import { GenerateActionBar } from "@/features/home/ui/GenerateActionBar";
 import { HistoryList } from "@/features/home/ui/HistoryList";
 import { HomeHeader } from "@/features/home/ui/HomeHeader";
 import { MaterialSelector } from "@/features/home/ui/MaterialSelector";
+import { TimelineEditor } from "@/features/home/ui/TimelineEditor";
+import { ClipTrimmer } from "@/features/home/ui/ClipTrimmer";
 import { PreviewPanel } from "@/features/home/ui/PreviewPanel";
 import { RefAudioPanel } from "@/features/home/ui/RefAudioPanel";
 import { ScriptEditor } from "@/features/home/ui/ScriptEditor";
 import { TitleSubtitlePanel } from "@/features/home/ui/TitleSubtitlePanel";
 import { VoiceSelector } from "@/features/home/ui/VoiceSelector";
+import { GeneratedAudiosPanel } from "@/features/home/ui/GeneratedAudiosPanel";

 export function HomePage() {
  const router = useRouter();
@@ -36,7 +40,6 @@ export function HomePage() {
    handleUpload,
    selectedMaterials,
    toggleMaterial,
-    reorderMaterials,
    handlePreviewMaterial,
    editingMaterialId,
    editMaterialName,
@@ -54,6 +57,9 @@ export function HomePage() {
    isTranslating,
    originalText,
    handleRestoreOriginal,
+    savedScripts,
+    handleSaveScript,
+    deleteSavedScript,
    showStylePreview,
    setShowStylePreview,
    videoTitle,
@@ -133,12 +139,40 @@ export function HomePage() {
    fetchGeneratedVideos,
    registerVideoRef,
    formatDate,
+    generatedAudios,
+    selectedAudio,
+    selectedAudioId,
+    isGeneratingAudio,
+    audioTask,
+    fetchGeneratedAudios,
+    handleGenerateAudio,
+    deleteAudio,
+    renameAudio,
+    selectAudio,
+    timelineSegments,
+    reorderSegments,
+    setSourceRange,
+    clipTrimmerOpen,
+    setClipTrimmerOpen,
+    clipTrimmerSegmentId,
+    setClipTrimmerSegmentId,
  } = useHomeController();

  useEffect(() => {
    router.prefetch("/publish");
  }, [router]);

+  const clipTrimmerSegment = useMemo(
+    () => timelineSegments.find((s) => s.id === clipTrimmerSegmentId) ?? null,
+    [timelineSegments, clipTrimmerSegmentId]
+  );
+
+  const clipTrimmerMaterialUrl = useMemo(() => {
+    if (!clipTrimmerSegment) return null;
+    const mat = materials.find((m) => m.id === clipTrimmerSegment.materialId);
+    return mat?.path ? resolveMediaUrl(mat.path) : null;
+  }, [clipTrimmerSegment, materials]);
+
  return (
    <div className="min-h-dvh">
      <HomeHeader />
@@ -147,34 +181,7 @@ export function HomePage() {
        <div className="grid grid-cols-1 lg:grid-cols-2 gap-8">
          {/* 左侧: 输入区域 */}
          <div className="space-y-6">
-            {/* 素材选择 */}
-            <MaterialSelector
-              materials={materials}
-              selectedMaterials={selectedMaterials}
-              isFetching={isFetching}
-              lastMaterialCount={lastMaterialCount}
-              editingMaterialId={editingMaterialId}
-              editMaterialName={editMaterialName}
-              isUploading={isUploading}
-              uploadProgress={uploadProgress}
-              uploadError={uploadError}
-              fetchError={fetchError}
-              apiBase={apiBase}
-              onUploadChange={handleUpload}
-              onRefresh={fetchMaterials}
-              onToggleMaterial={toggleMaterial}
-              onReorderMaterials={reorderMaterials}
-              onPreviewMaterial={handlePreviewMaterial}
-              onStartEditing={startMaterialEditing}
-              onEditNameChange={setEditMaterialName}
-              onSaveEditing={saveMaterialEditing}
-              onCancelEditing={cancelMaterialEditing}
-              onDeleteMaterial={deleteMaterial}
-              onClearUploadError={() => setUploadError(null)}
-              registerMaterialRef={registerMaterialRef}
-            />
-
-            {/* 文案输入 */}
+            {/* 1. 文案输入 */}
            <ScriptEditor
              text={text}
              onChangeText={setText}
@@ -185,9 +192,13 @@ export function HomePage() {
              isTranslating={isTranslating}
              hasOriginalText={originalText !== null}
              onRestoreOriginal={handleRestoreOriginal}
+              savedScripts={savedScripts}
+              onSaveScript={handleSaveScript}
+              onLoadScript={setText}
+              onDeleteScript={deleteSavedScript}
            />

-            {/* 标题和字幕设置 */}
+            {/* 2. 标题和字幕设置 */}
            <TitleSubtitlePanel
              showStylePreview={showStylePreview}
              onTogglePreview={() => setShowStylePreview((prev) => !prev)}
@@ -222,7 +233,7 @@ export function HomePage() {
              previewBaseHeight={materialDimensions?.height || 1920}
            />

-            {/* 配音方式选择 */}
+            {/* 3. 配音方式选择 */}
            <VoiceSelector
              ttsMode={ttsMode}
              onSelectTtsMode={setTtsMode}
@@ -260,7 +271,69 @@ export function HomePage() {
              )}
            />

-            {/* 背景音乐 */}
+            {/* 4. 配音列表 */}
+            <GeneratedAudiosPanel
+              generatedAudios={generatedAudios}
+              selectedAudioId={selectedAudioId}
+              isGeneratingAudio={isGeneratingAudio}
+              audioTask={audioTask}
+              onGenerateAudio={handleGenerateAudio}
+              onRefresh={() => fetchGeneratedAudios()}
+              onSelectAudio={selectAudio}
+              onDeleteAudio={deleteAudio}
+              onRenameAudio={renameAudio}
+              hasText={!!text.trim()}
+            />
+
+            {/* 5. 视频素材 */}
+            <MaterialSelector
+                materials={materials}
+                selectedMaterials={selectedMaterials}
+                isFetching={isFetching}
+                lastMaterialCount={lastMaterialCount}
+                editingMaterialId={editingMaterialId}
+                editMaterialName={editMaterialName}
+                isUploading={isUploading}
+                uploadProgress={uploadProgress}
+                uploadError={uploadError}
+                fetchError={fetchError}
+                apiBase={apiBase}
+                onUploadChange={handleUpload}
+                onRefresh={fetchMaterials}
+                onToggleMaterial={toggleMaterial}
+                onPreviewMaterial={handlePreviewMaterial}
+                onStartEditing={startMaterialEditing}
+                onEditNameChange={setEditMaterialName}
+                onSaveEditing={saveMaterialEditing}
+                onCancelEditing={cancelMaterialEditing}
+                onDeleteMaterial={deleteMaterial}
+                onClearUploadError={() => setUploadError(null)}
+                registerMaterialRef={registerMaterialRef}
+              />
+
+            {/* 5.5 时间轴编辑器 — 未选配音/素材时模糊遮挡 */}
+            <div className="relative">
+              {(!selectedAudio || selectedMaterials.length === 0) && (
+                <div className="absolute inset-0 bg-black/50 backdrop-blur-sm rounded-2xl flex items-center justify-center z-10">
+                  <p className="text-gray-400">
+                    {!selectedAudio ? "请先生成并选中配音" : "请先选择素材"}
+                  </p>
+                </div>
+              )}
+              <TimelineEditor
+                audioDuration={selectedAudio?.duration_sec ?? 0}
+                audioUrl={selectedAudio ? (resolveMediaUrl(selectedAudio.path) || "") : ""}
+                segments={timelineSegments}
+                materials={materials}
+                onReorderSegment={reorderSegments}
+                onClickSegment={(seg) => {
+                  setClipTrimmerSegmentId(seg.id);
+                  setClipTrimmerOpen(true);
+                }}
+              />
+            </div>
+
+            {/* 6. 背景音乐 */}
            <BgmPanel
              bgmList={bgmList}
              bgmLoading={bgmLoading}
@@ -278,12 +351,12 @@ export function HomePage() {
              registerBgmItemRef={registerBgmItemRef}
            />

-            {/* 生成按钮 */}
+            {/* 7. 生成按钮 */}
            <GenerateActionBar
              isGenerating={isGenerating}
              progress={currentTask?.progress || 0}
              materialCount={selectedMaterials.length}
-              disabled={isGenerating || selectedMaterials.length === 0 || (ttsMode === "voiceclone" && !selectedRefAudio)}
+              disabled={isGenerating || selectedMaterials.length === 0 || !selectedAudio}
              onGenerate={handleGenerate}
            />
          </div>
@@ -319,6 +392,19 @@ export function HomePage() {
        onClose={() => setExtractModalOpen(false)}
        onApply={(nextText) => setText(nextText)}
      />
+
+      <ClipTrimmer
+        isOpen={clipTrimmerOpen}
+        segment={clipTrimmerSegment}
+        materialUrl={clipTrimmerMaterialUrl}
+        onConfirm={(sourceStart, sourceEnd) => {
+          if (clipTrimmerSegmentId) {
+            setSourceRange(clipTrimmerSegmentId, sourceStart, sourceEnd);
+          }
+          setClipTrimmerOpen(false);
+        }}
+        onClose={() => setClipTrimmerOpen(false)}
+      />
    </div>
  );
 }
--- a/frontend/src/features/home/ui/MaterialSelector.tsx
+++ b/frontend/src/features/home/ui/MaterialSelector.tsx
@@ -1,21 +1,6 @@
 import { type ChangeEvent, type MouseEvent } from "react";
-import { Upload, RefreshCw, Eye, Trash2, X, Pencil, Check, GripVertical } from "lucide-react";
+import { Upload, RefreshCw, Eye, Trash2, X, Pencil, Check } from "lucide-react";
 import type { Material } from "@/shared/types/material";
-import {
-  DndContext,
-  closestCenter,
-  KeyboardSensor,
-  PointerSensor,
-  useSensor,
-  useSensors,
-  type DragEndEvent,
-} from "@dnd-kit/core";
-import {
-  SortableContext,
-  horizontalListSortingStrategy,
-  useSortable,
-} from "@dnd-kit/sortable";
-import { CSS } from "@dnd-kit/utilities";

 interface MaterialSelectorProps {
  materials: Material[];
@@ -32,7 +17,6 @@ interface MaterialSelectorProps {
  onUploadChange: (event: ChangeEvent<HTMLInputElement>) => void;
  onRefresh: () => void;
  onToggleMaterial: (id: string) => void;
-  onReorderMaterials: (activeId: string, overId: string) => void;
  onPreviewMaterial: (path: string) => void;
  onStartEditing: (material: Material, event: MouseEvent) => void;
  onEditNameChange: (value: string) => void;
@@ -43,61 +27,6 @@ interface MaterialSelectorProps {
  registerMaterialRef: (id: string, element: HTMLDivElement | null) => void;
 }

-function SortableChip({
-  id,
-  index,
-  label,
-  onRemove,
-}: {
-  id: string;
-  index: number;
-  label: string;
-  onRemove: () => void;
-}) {
-  const {
-    attributes,
-    listeners,
-    setNodeRef,
-    transform,
-    transition,
-    isDragging,
-  } = useSortable({ id });
-
-  const style = {
-    transform: CSS.Translate.toString(transform),
-    transition,
-  };
-
-  const circledNumbers = ["\u2460", "\u2461", "\u2462", "\u2463", "\u2464", "\u2465", "\u2466", "\u2467", "\u2468", "\u2469"];
-
-  return (
-    <div
-      ref={setNodeRef}
-      style={style}
-      className={`flex items-center gap-1 rounded-lg px-2 py-1 text-xs whitespace-nowrap transition-colors ${
-        isDragging
-          ? "bg-purple-500/50 border border-purple-400 text-white shadow-lg shadow-purple-500/30 z-10"
-          : "bg-purple-500/30 border border-purple-500/50 text-purple-200"
-      }`}
-    >
-      <span {...attributes} {...listeners} className="cursor-grab active:cursor-grabbing text-purple-400">
-        <GripVertical className="h-3 w-3" />
-      </span>
-      <span className="text-purple-300">{circledNumbers[index] || `${index + 1}`}</span>
-      <span className="max-w-[80px] truncate">{label}</span>
-      <button
-        onClick={(e) => {
-          e.stopPropagation();
-          onRemove();
-        }}
-        className="text-purple-400 hover:text-white ml-0.5"
-      >
-        <X className="h-3 w-3" />
-      </button>
-    </div>
-  );
-}
-
 export function MaterialSelector({
  materials,
  selectedMaterials,
@@ -113,7 +42,6 @@ export function MaterialSelector({
  onUploadChange,
  onRefresh,
  onToggleMaterial,
-  onReorderMaterials,
  onPreviewMaterial,
  onStartEditing,
  onEditNameChange,
@@ -123,21 +51,8 @@ export function MaterialSelector({
  onClearUploadError,
  registerMaterialRef,
 }: MaterialSelectorProps) {
-  const sensors = useSensors(
-    useSensor(PointerSensor, { activationConstraint: { distance: 5 } }),
-    useSensor(KeyboardSensor)
-  );
-
-  const handleDragEnd = (event: DragEndEvent) => {
-    const { active, over } = event;
-    if (over && active.id !== over.id) {
-      onReorderMaterials(String(active.id), String(over.id));
-    }
-  };
-
  const selectedSet = new Set(selectedMaterials);
  const isFull = selectedMaterials.length >= 4;
-  const circledNumbers = ["\u2460", "\u2461", "\u2462", "\u2463", "\u2464", "\u2465", "\u2466", "\u2467", "\u2468", "\u2469"];

  return (
    <div className="bg-white/5 rounded-2xl p-4 sm:p-6 border border-white/10 backdrop-blur-sm">
@@ -200,38 +115,6 @@ export function MaterialSelector({
        </div>
      )}

-      {/* 已选素材排列（拖拽排序区） - 仅当选中 >= 2 个时显示 */}
-      {selectedMaterials.length >= 2 && (
-        <div className="mb-3 p-3 bg-purple-500/10 rounded-xl border border-purple-500/20">
-          <div className="text-[11px] text-purple-300/70 mb-2">🎬 机位顺序 (拖拽调整)</div>
-          <DndContext
-            sensors={sensors}
-            collisionDetection={closestCenter}
-            onDragEnd={handleDragEnd}
-          >
-            <SortableContext
-              items={selectedMaterials}
-              strategy={horizontalListSortingStrategy}
-            >
-              <div className="flex flex-wrap gap-1.5">
-                {selectedMaterials.map((id, index) => {
-                  const m = materials.find((x) => x.id === id);
-                  return (
-                    <SortableChip
-                      key={id}
-                      id={id}
-                      index={index}
-                      label={m?.scene || m?.name || id}
-                      onRemove={() => onToggleMaterial(id)}
-                    />
-                  );
-                })}
-              </div>
-            </SortableContext>
-          </DndContext>
-        </div>
-      )}
-
      {fetchError ? (
        <div className="p-4 bg-red-500/20 text-red-200 rounded-xl text-sm mb-4">
          获取素材失败: {fetchError}
@@ -265,7 +148,6 @@ export function MaterialSelector({
        >
          {materials.map((m) => {
            const isSelected = selectedSet.has(m.id);
-            const selIndex = selectedMaterials.indexOf(m.id);
            return (
              <div
                key={m.id}
@@ -309,7 +191,7 @@ export function MaterialSelector({
                        : "border-white/30 text-transparent"
                        }`}
                    >
-                      {isSelected ? (selIndex >= 0 ? circledNumbers[selIndex] || "✓" : "✓") : ""}
+                      {isSelected ? "✓" : ""}
                    </span>
                    <div className="min-w-0">
                      <div className="text-white text-sm truncate">{m.scene || m.name}</div>
--- a/frontend/src/features/home/ui/ScriptEditor.tsx
+++ b/frontend/src/features/home/ui/ScriptEditor.tsx
@@ -1,5 +1,6 @@
 import { useEffect, useRef, useState } from "react";
-import { FileText, Languages, Loader2, RotateCcw, Sparkles } from "lucide-react";
+import { FileText, History, Languages, Loader2, RotateCcw, Save, Sparkles, Trash2 } from "lucide-react";
+import type { SavedScript } from "@/features/home/model/useSavedScripts";

 const LANGUAGES = [
  { code: "English", label: "英语 English" },
@@ -23,6 +24,10 @@ interface ScriptEditorProps {
  isTranslating: boolean;
  hasOriginalText: boolean;
  onRestoreOriginal: () => void;
+  savedScripts: SavedScript[];
+  onSaveScript: () => void;
+  onLoadScript: (content: string) => void;
+  onDeleteScript: (id: string) => void;
 }

 export function ScriptEditor({
@@ -35,9 +40,15 @@ export function ScriptEditor({
  isTranslating,
  hasOriginalText,
  onRestoreOriginal,
+  savedScripts,
+  onSaveScript,
+  onLoadScript,
+  onDeleteScript,
 }: ScriptEditorProps) {
  const [showLangMenu, setShowLangMenu] = useState(false);
  const langMenuRef = useRef<HTMLDivElement>(null);
+  const [showHistoryMenu, setShowHistoryMenu] = useState(false);
+  const historyMenuRef = useRef<HTMLDivElement>(null);

  useEffect(() => {
    if (!showLangMenu) return;
@@ -50,21 +61,81 @@ export function ScriptEditor({
    return () => document.removeEventListener("mousedown", handleClickOutside);
  }, [showLangMenu]);

+  useEffect(() => {
+    if (!showHistoryMenu) return;
+    const handleClickOutside = (e: MouseEvent) => {
+      if (historyMenuRef.current && !historyMenuRef.current.contains(e.target as Node)) {
+        setShowHistoryMenu(false);
+      }
+    };
+    document.addEventListener("mousedown", handleClickOutside);
+    return () => document.removeEventListener("mousedown", handleClickOutside);
+  }, [showHistoryMenu]);
+
  const handleSelectLang = (langCode: string) => {
    setShowLangMenu(false);
    onTranslate(langCode);
  };

+  const formatDate = (ts: number) => {
+    const d = new Date(ts);
+    return `${(d.getMonth() + 1).toString().padStart(2, "0")}-${d.getDate().toString().padStart(2, "0")} ${d.getHours().toString().padStart(2, "0")}:${d.getMinutes().toString().padStart(2, "0")}`;
+  };
+
  return (
    <div className="relative z-10 bg-white/5 rounded-2xl p-4 sm:p-6 border border-white/10 backdrop-blur-sm">
      <div className="mb-4 space-y-3">
        <h2 className="text-base sm:text-lg font-semibold text-white flex items-center gap-2">
          ✍️ 文案提取与编辑
        </h2>
-        <div className="flex gap-2 flex-wrap justify-end">
+        <div className="flex gap-2 flex-wrap justify-end items-center">
+          {/* 历史文案 */}
+          <div className="relative" ref={historyMenuRef}>
+            <button
+              onClick={() => setShowHistoryMenu((prev) => !prev)}
+              className="h-7 px-2.5 text-xs rounded transition-all whitespace-nowrap bg-gray-600 hover:bg-gray-500 text-white inline-flex items-center gap-1"
+            >
+              <History className="h-3.5 w-3.5" />
+              历史文案
+            </button>
+            {showHistoryMenu && (
+              <div className="absolute left-0 top-full mt-1 z-50 bg-gray-800 border border-white/10 rounded-lg shadow-xl py-1 min-w-[220px] max-h-[280px] overflow-y-auto">
+                {savedScripts.length === 0 ? (
+                  <div className="px-3 py-3 text-xs text-gray-500 text-center">暂无保存的文案</div>
+                ) : (
+                  savedScripts.map((script) => (
+                    <div
+                      key={script.id}
+                      className="flex items-center gap-1 px-3 py-1.5 hover:bg-white/10 transition-colors group"
+                    >
+                      <button
+                        onClick={() => {
+                          onLoadScript(script.content);
+                          setShowHistoryMenu(false);
+                        }}
+                        className="flex-1 text-left min-w-0"
+                      >
+                        <div className="text-xs text-gray-200 truncate">{script.name}</div>
+                        <div className="text-[10px] text-gray-500">{formatDate(script.savedAt)}</div>
+                      </button>
+                      <button
+                        onClick={(e) => {
+                          e.stopPropagation();
+                          onDeleteScript(script.id);
+                        }}
+                        className="opacity-0 group-hover:opacity-100 p-1 text-gray-500 hover:text-red-400 transition-all shrink-0"
+                      >
+                        <Trash2 className="h-3 w-3" />
+                      </button>
+                    </div>
+                  ))
+                )}
+              </div>
+            )}
+          </div>
          <button
            onClick={onOpenExtractModal}
-            className="px-2 py-1 text-xs rounded transition-all whitespace-nowrap bg-purple-600 hover:bg-purple-700 text-white flex items-center gap-1"
+            className="h-7 px-2.5 text-xs rounded transition-all whitespace-nowrap bg-purple-600 hover:bg-purple-700 text-white inline-flex items-center gap-1"
          >
            <FileText className="h-3.5 w-3.5" />
            文案提取助手
@@ -73,22 +144,22 @@ export function ScriptEditor({
            <button
              onClick={() => setShowLangMenu((prev) => !prev)}
              disabled={isTranslating || !text.trim()}
-              className={`px-2 py-1 text-xs rounded transition-all whitespace-nowrap ${
+              className={`h-7 px-2.5 text-xs rounded transition-all whitespace-nowrap inline-flex items-center gap-1 ${
                isTranslating || !text.trim()
                  ? "bg-gray-600 cursor-not-allowed text-gray-400"
                  : "bg-gradient-to-r from-emerald-600 to-teal-600 hover:from-emerald-700 hover:to-teal-700 text-white"
              }`}
            >
              {isTranslating ? (
-                <span className="flex items-center gap-1">
+                <>
                  <Loader2 className="h-3.5 w-3.5 animate-spin" />
                  翻译中...
-                </span>
+                </>
              ) : (
-                <span className="flex items-center gap-1">
+                <>
                  <Languages className="h-3.5 w-3.5" />
                  AI多语言
-                </span>
+                </>
              )}
            </button>
            {showLangMenu && (
@@ -120,21 +191,21 @@ export function ScriptEditor({
          <button
            onClick={onGenerateMeta}
            disabled={isGeneratingMeta || !text.trim()}
-            className={`px-2 py-1 text-xs rounded transition-all whitespace-nowrap ${isGeneratingMeta || !text.trim()
+            className={`h-7 px-2.5 text-xs rounded transition-all whitespace-nowrap inline-flex items-center gap-1 ${isGeneratingMeta || !text.trim()
              ? "bg-gray-600 cursor-not-allowed text-gray-400"
              : "bg-gradient-to-r from-blue-600 to-cyan-600 hover:from-blue-700 hover:to-cyan-700 text-white"
              }`}
          >
            {isGeneratingMeta ? (
-              <span className="flex items-center gap-1">
+              <>
                <Loader2 className="h-3.5 w-3.5 animate-spin" />
                生成中...
-              </span>
+              </>
            ) : (
-              <span className="flex items-center gap-1">
+              <>
                <Sparkles className="h-3.5 w-3.5" />
                AI生成标题标签
-              </span>
+              </>
            )}
          </button>
        </div>
@@ -145,9 +216,20 @@ export function ScriptEditor({
        placeholder="请输入你想说的话..."
        className="w-full h-40 bg-black/30 border border-white/10 rounded-xl p-4 text-white placeholder-gray-500 resize-none focus:outline-none focus:border-purple-500 transition-colors hide-scrollbar"
      />
-      <div className="flex justify-between mt-2 text-sm text-gray-400">
+      <div className="flex items-center justify-between mt-2 text-sm text-gray-400">
        <span>{text.length} 字</span>
-        <span>预计时长: ~{Math.ceil(text.length / 4)} 秒</span>
+        <button
+          onClick={onSaveScript}
+          disabled={!text.trim()}
+          className={`px-2.5 py-1 text-xs rounded transition-all flex items-center gap-1 ${
+            !text.trim()
+              ? "bg-gray-700 cursor-not-allowed text-gray-500"
+              : "bg-amber-600/80 hover:bg-amber-600 text-white"
+          }`}
+        >
+          <Save className="h-3 w-3" />
+          保存文案
+        </button>
      </div>
    </div>
  );
--- a/frontend/src/features/home/ui/TimelineEditor.tsx
+++ b/frontend/src/features/home/ui/TimelineEditor.tsx
@@ -0,0 +1,283 @@
+import { useEffect, useRef, useCallback, useState } from "react";
+import WaveSurfer from "wavesurfer.js";
+import type { TimelineSegment } from "@/features/home/model/useTimelineEditor";
+import type { Material } from "@/shared/types/material";
+
+interface TimelineEditorProps {
+  audioDuration: number;
+  audioUrl: string;
+  segments: TimelineSegment[];
+  materials: Material[];
+  onReorderSegment: (fromIdx: number, toIdx: number) => void;
+  onClickSegment: (segment: TimelineSegment) => void;
+}
+
+function formatTime(sec: number): string {
+  const m = Math.floor(sec / 60);
+  const s = sec % 60;
+  return `${String(m).padStart(2, "0")}:${s.toFixed(1).padStart(4, "0")}`;
+}
+
+export function TimelineEditor({
+  audioDuration,
+  audioUrl,
+  segments,
+  materials,
+  onReorderSegment,
+  onClickSegment,
+}: TimelineEditorProps) {
+  const waveRef = useRef<HTMLDivElement>(null);
+  const wsRef = useRef<WaveSurfer | null>(null);
+  const [waveReady, setWaveReady] = useState(false);
+  const [isPlaying, setIsPlaying] = useState(false);
+
+  // Refs for high-frequency DOM updates (avoid 60fps re-renders)
+  const playheadRef = useRef<HTMLDivElement>(null);
+  const timeRef = useRef<HTMLSpanElement>(null);
+  const audioDurationRef = useRef(audioDuration);
+  audioDurationRef.current = audioDuration;
+
+  // Drag-to-reorder state
+  const [dragFromIdx, setDragFromIdx] = useState<number | null>(null);
+  const [dragOverIdx, setDragOverIdx] = useState<number | null>(null);
+
+  // Create / recreate wavesurfer when audioUrl changes
+  useEffect(() => {
+    if (!waveRef.current || !audioUrl) return;
+
+    // Destroy previous instance
+    if (wsRef.current) {
+      wsRef.current.destroy();
+      wsRef.current = null;
+    }
+
+    const ws = WaveSurfer.create({
+      container: waveRef.current,
+      height: 56,
+      waveColor: "#6d28d9",
+      progressColor: "#a855f7",
+      barWidth: 2,
+      barGap: 1,
+      barRadius: 2,
+      cursorWidth: 1,
+      cursorColor: "#e879f9",
+      interact: true,
+      normalize: true,
+    });
+
+    // Click waveform → seek + auto-play
+    ws.on("interaction", () => ws.play());
+    ws.on("play", () => setIsPlaying(true));
+    ws.on("pause", () => setIsPlaying(false));
+    ws.on("finish", () => {
+      setIsPlaying(false);
+      if (playheadRef.current) playheadRef.current.style.display = "none";
+    });
+    // High-frequency: update playhead + time via refs (no React re-render)
+    ws.on("timeupdate", (time: number) => {
+      const dur = audioDurationRef.current;
+      if (playheadRef.current && dur > 0) {
+        playheadRef.current.style.left = `${(time / dur) * 100}%`;
+        playheadRef.current.style.display = "block";
+      }
+      if (timeRef.current) {
+        timeRef.current.textContent = formatTime(time);
+      }
+    });
+
+    ws.load(audioUrl);
+    wsRef.current = ws;
+
+    return () => {
+      ws.destroy();
+      wsRef.current = null;
+      setIsPlaying(false);
+      if (playheadRef.current) playheadRef.current.style.display = "none";
+      if (timeRef.current) timeRef.current.textContent = formatTime(0);
+    };
+  }, [audioUrl, waveReady]);
+
+  // Callback ref to detect when waveRef div mounts
+  const waveCallbackRef = useCallback((node: HTMLDivElement | null) => {
+    (waveRef as React.MutableRefObject<HTMLDivElement | null>).current = node;
+    setWaveReady(!!node);
+  }, []);
+
+  const handlePlayPause = useCallback(() => {
+    wsRef.current?.playPause();
+  }, []);
+
+  // Drag-to-reorder handlers
+  const handleDragStart = useCallback((idx: number, e: React.DragEvent) => {
+    setDragFromIdx(idx);
+    e.dataTransfer.effectAllowed = "move";
+    e.dataTransfer.setData("text/plain", String(idx));
+  }, []);
+
+  const handleDragOver = useCallback((idx: number, e: React.DragEvent) => {
+    e.preventDefault();
+    e.dataTransfer.dropEffect = "move";
+    setDragOverIdx(idx);
+  }, []);
+
+  const handleDragLeave = useCallback(() => {
+    setDragOverIdx(null);
+  }, []);
+
+  const handleDrop = useCallback((toIdx: number, e: React.DragEvent) => {
+    e.preventDefault();
+    const fromIdx = parseInt(e.dataTransfer.getData("text/plain"), 10);
+    if (!isNaN(fromIdx) && fromIdx !== toIdx) {
+      onReorderSegment(fromIdx, toIdx);
+    }
+    setDragFromIdx(null);
+    setDragOverIdx(null);
+  }, [onReorderSegment]);
+
+  const handleDragEnd = useCallback(() => {
+    setDragFromIdx(null);
+    setDragOverIdx(null);
+  }, []);
+
+  // Filter visible vs overflow segments
+  const visibleSegments = segments.filter((s) => s.start < audioDuration);
+  const overflowSegments = segments.filter((s) => s.start >= audioDuration);
+  const hasSegments = visibleSegments.length > 0;
+
+  return (
+    <div className="bg-white/5 rounded-2xl p-4 sm:p-6 border border-white/10 backdrop-blur-sm">
+      <div className="flex items-center justify-between mb-3">
+        <h2 className="text-base sm:text-lg font-semibold text-white flex items-center gap-2">
+          🎞️ 时间轴编辑
+        </h2>
+        {audioUrl && (
+          <div className="flex items-center gap-2 text-xs text-gray-400">
+            <button
+              onClick={handlePlayPause}
+              className="w-7 h-7 flex items-center justify-center rounded-full bg-white/10 hover:bg-white/20 text-white transition-colors"
+              title={isPlaying ? "暂停" : "播放"}
+            >
+              {isPlaying ? "⏸" : "▶"}
+            </button>
+            <span ref={timeRef} className="tabular-nums">00:00.0</span>
+            <span className="text-gray-600">/</span>
+            <span className="tabular-nums">{formatTime(audioDuration)}</span>
+          </div>
+        )}
+      </div>
+
+      {/* Waveform — always rendered so ref stays mounted */}
+      <div className="relative mb-1">
+        <div ref={waveCallbackRef} className="rounded-lg overflow-hidden bg-black/20 cursor-pointer" style={{ minHeight: 56 }} />
+      </div>
+
+      {/* Segment blocks or empty placeholder */}
+      {hasSegments ? (
+        <>
+          <div className="relative h-14 flex select-none">
+            {/* Playhead — syncs with audio playback */}
+            <div
+              ref={playheadRef}
+              className="absolute top-0 h-full w-0.5 bg-fuchsia-400 z-10 pointer-events-none"
+              style={{ display: "none", left: "0%" }}
+            />
+            {visibleSegments.map((seg, i) => {
+              const left = (seg.start / audioDuration) * 100;
+              const width = ((seg.end - seg.start) / audioDuration) * 100;
+              const segDur = seg.end - seg.start;
+              const isDragTarget = dragOverIdx === i && dragFromIdx !== i;
+
+              // Compute loop portion for the last visible segment
+              const isLastVisible = i === visibleSegments.length - 1;
+              let loopPercent = 0;
+              if (isLastVisible && audioDuration > 0) {
+                const mat = materials.find((m) => m.id === seg.materialId);
+                const matDur = mat?.duration_sec ?? 0;
+                const effDur = (seg.sourceEnd > seg.sourceStart)
+                  ? (seg.sourceEnd - seg.sourceStart)
+                  : matDur;
+                if (effDur > 0 && segDur > effDur + 0.1) {
+                  loopPercent = ((segDur - effDur) / segDur) * 100;
+                }
+              }
+
+              return (
+                <div key={seg.id} className="absolute top-0 h-full" style={{ left: `${left}%`, width: `${width}%` }}>
+                  <button
+                    draggable
+                    onDragStart={(e) => handleDragStart(i, e)}
+                    onDragOver={(e) => handleDragOver(i, e)}
+                    onDragLeave={handleDragLeave}
+                    onDrop={(e) => handleDrop(i, e)}
+                    onDragEnd={handleDragEnd}
+                    onClick={() => onClickSegment(seg)}
+                    className={`relative w-full h-full rounded-lg flex flex-col items-center justify-center overflow-hidden cursor-grab active:cursor-grabbing transition-all border ${
+                      isDragTarget
+                        ? "ring-2 ring-purple-400 border-purple-400 scale-[1.02]"
+                        : dragFromIdx === i
+                        ? "opacity-50 border-white/10"
+                        : "hover:opacity-90 border-white/10"
+                    }`}
+                    style={{ backgroundColor: seg.color + "33", borderColor: isDragTarget ? undefined : seg.color + "66" }}
+                    title={`拖拽可调换顺序 · 点击设置截取范围\n${seg.materialName}\n${segDur.toFixed(1)}s${loopPercent > 0 ? ` (含循环 ${(segDur * loopPercent / 100).toFixed(1)}s)` : ""}`}
+                  >
+                    <span className="text-[11px] text-white/90 truncate max-w-full px-1 leading-tight z-[1]">
+                      {seg.materialName}
+                    </span>
+                    <span className="text-[10px] text-white/60 leading-tight z-[1]">
+                      {segDur.toFixed(1)}s
+                    </span>
+                    {seg.sourceStart > 0 && (
+                      <span className="text-[9px] text-amber-400/80 leading-tight z-[1]">
+                        ✂ {seg.sourceStart.toFixed(1)}s
+                      </span>
+                    )}
+                    {/* Loop fill stripe overlay */}
+                    {loopPercent > 0 && (
+                      <div
+                        className="absolute top-0 right-0 h-full pointer-events-none flex items-center justify-center"
+                        style={{
+                          width: `${loopPercent}%`,
+                          background: `repeating-linear-gradient(-45deg, transparent, transparent 3px, rgba(255,255,255,0.07) 3px, rgba(255,255,255,0.07) 6px)`,
+                          borderLeft: "1px dashed rgba(255,255,255,0.25)",
+                        }}
+                      >
+                        <span className="text-[9px] text-white/30">循环</span>
+                      </div>
+                    )}
+                  </button>
+                </div>
+              );
+            })}
+          </div>
+
+          {/* Overflow segments — shown as gray chips */}
+          {overflowSegments.length > 0 && (
+            <div className="flex flex-wrap items-center gap-1.5 mt-1.5">
+              <span className="text-[10px] text-gray-500">未使用:</span>
+              {overflowSegments.map((seg) => (
+                <span
+                  key={seg.id}
+                  className="text-[10px] text-gray-500 bg-white/5 border border-white/10 rounded px-1.5 py-0.5"
+                >
+                  {seg.materialName}
+                </span>
+              ))}
+            </div>
+          )}
+
+          <p className="text-[10px] text-gray-500 mt-1.5">
+            点击波形定位播放 · 拖拽色块调换顺序 · 点击色块设置截取范围
+          </p>
+        </>
+      ) : (
+        <>
+          <div className="h-14 bg-white/5 rounded-lg" />
+          <p className="text-[10px] text-gray-500 mt-1.5">
+            选中配音和素材后可编辑时间轴
+          </p>
+        </>
+      )}
+    </div>
+  );
+}
--- a/frontend/src/shared/types/material.ts
+++ b/frontend/src/shared/types/material.ts
@@ -4,4 +4,5 @@ export interface Material {
  path: string;
  size_mb: number;
  scene?: string;
+  duration_sec?: number;
 }
--- a/models/Qwen3-TTS/qwen_tts_server.py
+++ b/models/Qwen3-TTS/qwen_tts_server.py
@@ -120,6 +120,7 @@ async def generate(
    if not _model_loaded:
        raise HTTPException(status_code=503, detail="Model not loaded")

+    import torch
    import soundfile as sf

    # 保存上传的参考音频到临时文件
@@ -132,7 +133,7 @@ async def generate(
    output_path = tempfile.mktemp(suffix=".wav")

    try:
-        print(f"🎤 Generating: {text[:30]}...")
+        print(f"🎤 Generating: {text[:50]}... ({len(text)} chars)")
        print(f"📝 Ref text: {ref_text[:50]}...")
        print(f"🌐 Language: {language}")

@@ -148,6 +149,9 @@ async def generate(
            ref_text=ref_text,
        )

+        # 释放 CUDA 缓存，防止显存碎片累积
+        torch.cuda.empty_cache()
+
        sf.write(output_path, wavs[0], sr)

        duration = len(wavs[0]) / sr
@@ -158,11 +162,17 @@ async def generate(
            output_path,
            media_type="audio/wav",
            filename="output.wav",
-            background=None  # 让客户端下载完再删除
+            background=None
        )

    except Exception as e:
        print(f"❌ Generation failed: {e}")
+        # 释放 CUDA 缓存
+        try:
+            import torch
+            torch.cuda.empty_cache()
+        except:
+            pass
        raise HTTPException(status_code=500, detail=str(e))
    finally:
        # 清理参考音频临时文件
--- a/run_qwen_tts.sh
+++ b/run_qwen_tts.sh
@@ -5,5 +5,7 @@

 cd /home/rongye/ProgramFiles/ViGent2/models/Qwen3-TTS

-# 使用 qwen-tts conda 环境的 Python
-/home/rongye/ProgramFiles/miniconda3/envs/qwen-tts/bin/python qwen_tts_server.py
+# 确保 conda env 的 bin 目录在 PATH 中，让 sox 等工具可被找到
+export PATH="/home/rongye/ProgramFiles/miniconda3/envs/qwen-tts/bin:$PATH"
+
+python qwen_tts_server.py