diff --git a/docs/en/getting-started/configuration.md b/docs/en/getting-started/configuration.md
index bf37a14..fffe0e9 100644
--- a/docs/en/getting-started/configuration.md
+++ b/docs/en/getting-started/configuration.md
@@ -24,25 +24,29 @@ LLM (Large Language Model) is used to generate video scripts.
 
 ---
 
-## Image Configuration
+## Image/Video Generation Configuration
 
 Two options available:
 
-### Local Deployment (Recommended)
+### Local Deployment
 
 Using local ComfyUI service:
 
 1. Install and start ComfyUI
 2. Enter ComfyUI URL (default `http://127.0.0.1:8188`)
 3. Click "Test Connection" to verify
+4. (Optional) Enter ComfyUI API Key (get from [Comfy Platform](https://platform.comfy.org/profile/api-keys))
 
-### Cloud Deployment
+### Cloud Deployment (Recommended)
 
-Using RunningHub cloud service:
+Using RunningHub cloud service, no local GPU required:
 
 1. Register for a RunningHub account
 2. Obtain API Key
 3. Enter API Key in configuration
+4. Configure advanced options (optional):
+   - **Concurrent Limit**: Set number of simultaneous tasks (1-10, default 1 for regular members)
+   - **Instance Type**: Choose 24GB or 48GB VRAM machine (48GB for large video generation)
 
 ---
 
diff --git a/docs/en/reference/api-overview.md b/docs/en/reference/api-overview.md
index 057bb2a..abdc6b7 100644
--- a/docs/en/reference/api-overview.md
+++ b/docs/en/reference/api-overview.md
@@ -1,10 +1,10 @@
 # API Overview
 
-Pixelle-Video Python API reference documentation.
+Pixelle-Video provides both Python SDK and HTTP REST API.
 
 ---
 
-## Core Classes
+## Python SDK
 
 ### PixelleVideoCore
 
@@ -17,10 +17,6 @@ pixelle = PixelleVideoCore()
 await pixelle.initialize()
 ```
 
----
-
-## Main Methods
-
 ### generate_video()
 
 Primary method for generating videos.
@@ -32,21 +28,114 @@ Primary method for generating videos.
 - `n_scenes` (int): Number of scenes
 - `title` (str, optional): Video title
 - `tts_workflow` (str): TTS workflow
-- `image_workflow` (str): Image generation workflow
+- `media_workflow` (str): Media generation workflow (image or video)
 - `frame_template` (str): Video template
+- `template_params` (dict, optional): Custom template parameters
 - `bgm_path` (str, optional): BGM file path
+- `bgm_volume` (float): BGM volume (0.0-1.0)
 
 **Returns**: `VideoResult` object
 
 ---
 
-## Examples
+## HTTP REST API
 
-Check the `examples/` directory for more examples.
+Start the API server:
+
+```bash
+uv run uvicorn api.app:app --host 0.0.0.0 --port 8000
+```
+
+### Video Generation - Synchronous
+
+`POST /api/video/generate/sync`
+
+Generate video synchronously, waits until completion. Suitable for small videos (< 30 seconds).
+
+**Request Body**:
+
+```json
+{
+  "text": "Why you should develop a reading habit",
+  "mode": "generate",
+  "n_scenes": 5,
+  "frame_template": "1080x1920/image_default.html",
+  "template_params": {
+    "accent_color": "#3498db",
+    "background": "https://example.com/custom-bg.jpg"
+  },
+  "title": "The Power of Reading"
+}
+```
+
+**Response**:
+
+```json
+{
+  "success": true,
+  "message": "Success",
+  "video_url": "http://localhost:8000/api/files/xxx/final.mp4",
+  "duration": 45.5,
+  "file_size": 12345678
+}
+```
+
+### Video Generation - Asynchronous
+
+`POST /api/video/generate/async`
+
+Generate video asynchronously, returns task ID immediately. Suitable for large videos.
+
+**Response**:
+
+```json
+{
+  "success": true,
+  "message": "Task created successfully",
+  "task_id": "abc123"
+}
+```
+
+### Query Task Status
+
+`GET /api/tasks/{task_id}`
+
+**Response**:
+
+```json
+{
+  "task_id": "abc123",
+  "status": "completed",
+  "result": {
+    "video_url": "http://localhost:8000/api/files/xxx/final.mp4",
+    "duration": 45.5,
+    "file_size": 12345678
+  }
+}
+```
+
+---
+
+## Request Parameters
+
+| Parameter | Type | Required | Description |
+|-----------|------|----------|-------------|
+| `text` | string | Yes | Topic or complete script |
+| `mode` | string | No | `"generate"` (AI generates) or `"fixed"` (use text as-is) |
+| `n_scenes` | int | No | Number of scenes (1-20), only used in generate mode |
+| `title` | string | No | Video title (auto-generated if not provided) |
+| `frame_template` | string | No | Template path, e.g., `1080x1920/image_default.html` |
+| `template_params` | object | No | Custom template parameters (colors, backgrounds, etc.) |
+| `media_workflow` | string | No | Media workflow (image or video generation) |
+| `tts_workflow` | string | No | TTS workflow |
+| `ref_audio` | string | No | Reference audio path for voice cloning |
+| `prompt_prefix` | string | No | Image style prefix |
+| `bgm_path` | string | No | BGM file path |
+| `bgm_volume` | float | No | BGM volume (0.0-1.0, default 0.3) |
 
 ---
 
 ## More Information
 
-Detailed API documentation coming soon.
+API documentation is also available via Swagger UI: `http://localhost:8000/docs`
 
diff --git a/docs/en/reference/config-schema.md b/docs/en/reference/config-schema.md
index 60767dd..b822a80 100644
--- a/docs/en/reference/config-schema.md
+++ b/docs/en/reference/config-schema.md
@@ -8,14 +8,16 @@ Detailed explanation of the `config.yaml` configuration file.
 
 ```yaml
 llm:
-  provider: openai
   api_key: "your-api-key"
   base_url: "https://dashscope.aliyuncs.com/compatible-mode/v1"
   model: "qwen-plus"
 
 comfyui:
   comfyui_url: "http://127.0.0.1:8188"
+  comfyui_api_key: ""  # ComfyUI API key (optional)
   runninghub_api_key: ""
+  runninghub_concurrent_limit: 1  # Concurrent limit (1-10)
+  runninghub_instance_type: ""  # Instance type (optional, set to "plus" for 48GB VRAM)
   
   image:
     default_workflow: "runninghub/image_flux.json"
@@ -27,15 +29,17 @@ comfyui:
   
   tts:
     default_workflow: "selfhost/tts_edge.json"
+
+template:
+  default_template: "1080x1920/image_default.html"
 ```
 
 ---
 
 ## LLM Configuration
 
-- `provider`: Provider (currently only supports openai-compatible interfaces)
 - `api_key`: API key
-- `base_url`: API service address
+- `base_url`: API service address (supports any OpenAI-compatible interface)
 - `model`: Model name
 
 ---
@@ -44,8 +48,16 @@ comfyui:
 
 ### Basic Configuration
 
-- `comfyui_url`: Local ComfyUI address
-- `runninghub_api_key`: RunningHub API key (optional)
+- `comfyui_url`: Local ComfyUI address (default `http://127.0.0.1:8188`)
+- `comfyui_api_key`: ComfyUI API key (optional, for [Comfy Platform](https://platform.comfy.org/profile/api-keys))
+
+### RunningHub Cloud Configuration
+
+- `runninghub_api_key`: RunningHub API key (required for cloud workflows)
+- `runninghub_concurrent_limit`: Concurrent execution limit (1-10, default 1 for regular members)
+- `runninghub_instance_type`: Instance type (optional)
+  - Empty or unset: Use 24GB VRAM machine
+  - `"plus"`: Use 48GB VRAM machine (suitable for large video generation)
 
 ### Image Configuration
 
@@ -65,6 +77,12 @@ comfyui:
 
 ---
 
+## Template Configuration
+
+- `default_template`: Default frame template path (e.g., `1080x1920/image_default.html`)
+
+---
+
 ## More Information
 
 The configuration file is automatically created on first run.
diff --git a/docs/en/user-guide/web-ui.md b/docs/en/user-guide/web-ui.md
index 76846af..844fcf9 100644
--- a/docs/en/user-guide/web-ui.md
+++ b/docs/en/user-guide/web-ui.md
@@ -11,6 +11,7 @@ The Web interface uses a three-column layout:
 - **Left Panel**: Content input and audio settings
 - **Middle Panel**: Voice and visual settings  
 - **Right Panel**: Video generation and preview
+- **Sidebar**: System configuration and FAQ
 
 ---
 
@@ -27,6 +28,14 @@ First-time use requires configuring LLM and image generation services. See [Conf
 - **AI Generate Content**: Enter a topic, AI creates script automatically
 - **Fixed Script Content**: Enter complete script directly
 
+### Fixed Script Split Mode
+
+When using fixed script mode, you can choose how to split the content:
+
+- **By Paragraph**: Split by empty lines, each paragraph becomes a scene
+- **By Line**: Split by line breaks, each line becomes a scene
+- **By Sentence**: Smart sentence boundary detection, each sentence becomes a scene
+
 ### Background Music
 
 - Built-in music supported
@@ -50,17 +59,19 @@ First-time use requires configuring LLM and image generation services. See [Conf
 
 ## Visual Settings
 
-### Image Generation
+### Image/Video Generation
 
-- Select image generation workflow
-- Set image dimensions
+- Select media generation workflow (image or video)
 - Adjust prompt prefix to control style
 
 ### Video Template
 
-- Choose video template
-- Supports portrait/landscape/square
-- Preview template effects
+- **Template Preview Gallery**: Visually preview all available templates
+- Supports portrait (1080x1920) / landscape (1920x1080) / square (1080x1080)
+- Template types:
+  - `static_*.html`: Static templates (no AI-generated media)
+  - `image_*.html`: Image templates (requires AI-generated images)
+  - `video_*.html`: Video templates (requires AI-generated videos)
 
 ---
 
@@ -69,9 +80,19 @@ First-time use requires configuring LLM and image generation services. See [Conf
 After clicking "Generate Video", the system will:
 
 1. Generate video script
-2. Generate images for each scene
+2. Generate images/videos for each scene
 3. Synthesize voice narration
 4. Compose final video
 
 Automatically previews when complete.
 
+---
+
+## FAQ
+
+The sidebar includes built-in FAQ for quick reference:
+
+- Common configuration issues
+- Generation failure solutions
+- Performance optimization tips
+
diff --git a/docs/zh/getting-started/configuration.md b/docs/zh/getting-started/configuration.md
index 76a4875..6f872e5 100644
--- a/docs/zh/getting-started/configuration.md
+++ b/docs/zh/getting-started/configuration.md
@@ -24,25 +24,29 @@ LLM（大语言模型）用于生成视频文案。
 
 ---
 
-## 图像配置
+## 图像/视频生成配置
 
 支持两种方式：
 
-### 本地部署（推荐）
+### 本地部署
 
 使用本地 ComfyUI 服务：
 
 1. 安装并启动 ComfyUI
 2. 填写 ComfyUI URL（默认 `http://127.0.0.1:8188`）
 3. 点击「测试连接」确认服务可用
+4. （可选）填写 ComfyUI API Key（从 [Comfy Platform](https://platform.comfy.org/profile/api-keys) 获取）
 
-### 云端部署
+### 云端部署（推荐）
 
-使用 RunningHub 云端服务：
+使用 RunningHub 云端服务，无需本地 GPU：
 
 1. 注册 RunningHub 账号
 2. 获取 API Key
 3. 在配置中填写 API Key
+4. 配置高级选项（可选）：
+   - **并发限制**: 设置同时执行的任务数（1-10，普通会员默认为 1）
+   - **实例类型**: 选择 24GB 或 48GB 显存机器（48GB 适合大尺寸视频生成）
 
 ---
 
diff --git a/docs/zh/reference/api-overview.md b/docs/zh/reference/api-overview.md
index 2939719..bcdd467 100644
--- a/docs/zh/reference/api-overview.md
+++ b/docs/zh/reference/api-overview.md
@@ -1,10 +1,10 @@
 # API 概览
 
-Pixelle-Video Python API 参考文档。
+Pixelle-Video 提供 Python SDK 和 HTTP REST API 两种方式。
 
 ---
 
-## 核心类
+## Python SDK
 
 ### PixelleVideoCore
 
@@ -17,10 +17,6 @@ pixelle = PixelleVideoCore()
 await pixelle.initialize()
 ```
 
----
-
-## 主要方法
-
 ### generate_video()
 
 生成视频的主要方法。
@@ -32,21 +28,114 @@ await pixelle.initialize()
 - `n_scenes` (int): 分镜数量
 - `title` (str, optional): 视频标题
 - `tts_workflow` (str): TTS 工作流
-- `image_workflow` (str): 图像生成工作流
+- `media_workflow` (str): 媒体生成工作流（图像或视频）
 - `frame_template` (str): 视频模板
+- `template_params` (dict, optional): 模板自定义参数
 - `bgm_path` (str, optional): BGM 文件路径
+- `bgm_volume` (float): BGM 音量 (0.0-1.0)
 
 **返回**: `VideoResult` 对象
 
 ---
 
-## 示例
+## HTTP REST API
 
-查看 `examples/` 目录获取更多示例。
+启动 API 服务器：
+
+```bash
+uv run uvicorn api.app:app --host 0.0.0.0 --port 8000
+```
+
+### 视频生成 - 同步
+
+`POST /api/video/generate/sync`
+
+同步生成视频，等待完成后返回结果。适合小视频（< 30 秒）。
+
+**请求体**:
+
+```json
+{
+  "text": "为什么要养成阅读习惯",
+  "mode": "generate",
+  "n_scenes": 5,
+  "frame_template": "1080x1920/image_default.html",
+  "template_params": {
+    "accent_color": "#3498db",
+    "background": "https://example.com/custom-bg.jpg"
+  },
+  "title": "阅读的力量"
+}
+```
+
+**响应**:
+
+```json
+{
+  "success": true,
+  "message": "Success",
+  "video_url": "http://localhost:8000/api/files/xxx/final.mp4",
+  "duration": 45.5,
+  "file_size": 12345678
+}
+```
+
+### 视频生成 - 异步
+
+`POST /api/video/generate/async`
+
+异步生成视频，立即返回任务 ID。适合大视频。
+
+**响应**:
+
+```json
+{
+  "success": true,
+  "message": "Task created successfully",
+  "task_id": "abc123"
+}
+```
+
+### 查询任务状态
+
+`GET /api/tasks/{task_id}`
+
+**响应**:
+
+```json
+{
+  "task_id": "abc123",
+  "status": "completed",
+  "result": {
+    "video_url": "http://localhost:8000/api/files/xxx/final.mp4",
+    "duration": 45.5,
+    "file_size": 12345678
+  }
+}
+```
+
+---
+
+## 请求参数说明
+
+| 参数 | 类型 | 必填 | 说明 |
+|------|------|------|------|
+| `text` | string | 是 | 主题或完整文案 |
+| `mode` | string | 否 | `"generate"` (AI 生成) 或 `"fixed"` (固定文案) |
+| `n_scenes` | int | 否 | 分镜数量 (1-20)，仅 generate 模式有效 |
+| `title` | string | 否 | 视频标题（不填则自动生成） |
+| `frame_template` | string | 否 | 模板路径，如 `1080x1920/image_default.html` |
+| `template_params` | object | 否 | 模板自定义参数（颜色、背景等） |
+| `media_workflow` | string | 否 | 媒体工作流（图像或视频生成） |
+| `tts_workflow` | string | 否 | TTS 工作流 |
+| `ref_audio` | string | 否 | 声音克隆参考音频路径 |
+| `prompt_prefix` | string | 否 | 图像风格前缀 |
+| `bgm_path` | string | 否 | BGM 文件路径 |
+| `bgm_volume` | float | 否 | BGM 音量 (0.0-1.0，默认 0.3) |
 
 ---
 
 ## 更多信息
 
-详细的 API 文档即将推出。
+API 文档也可通过 Swagger UI 访问：`http://localhost:8000/docs`
 
diff --git a/docs/zh/reference/config-schema.md b/docs/zh/reference/config-schema.md
index fb6c8f5..0b93d73 100644
--- a/docs/zh/reference/config-schema.md
+++ b/docs/zh/reference/config-schema.md
@@ -8,14 +8,16 @@
 
 ```yaml
 llm:
-  provider: openai
   api_key: "your-api-key"
   base_url: "https://dashscope.aliyuncs.com/compatible-mode/v1"
   model: "qwen-plus"
 
 comfyui:
   comfyui_url: "http://127.0.0.1:8188"
+  comfyui_api_key: ""  # ComfyUI API 密钥（可选）
   runninghub_api_key: ""
+  runninghub_concurrent_limit: 1  # 并发限制 (1-10)
+  runninghub_instance_type: ""  # 实例类型（可选，设为 "plus" 使用 48GB 显存）
   
   image:
     default_workflow: "runninghub/image_flux.json"
@@ -27,15 +29,17 @@ comfyui:
   
   tts:
     default_workflow: "selfhost/tts_edge.json"
+
+template:
+  default_template: "1080x1920/image_default.html"
 ```
 
 ---
 
 ## LLM 配置
 
-- `provider`: 提供商（目前仅支持 openai 兼容接口）
 - `api_key`: API 密钥
-- `base_url`: API 服务地址
+- `base_url`: API 服务地址（支持任何 OpenAI 兼容接口）
 - `model`: 模型名称
 
 ---
@@ -44,8 +48,16 @@ comfyui:
 
 ### 基础配置
 
-- `comfyui_url`: 本地 ComfyUI 地址
-- `runninghub_api_key`: RunningHub API 密钥（可选）
+- `comfyui_url`: 本地 ComfyUI 地址（默认 `http://127.0.0.1:8188`）
+- `comfyui_api_key`: ComfyUI API 密钥（可选，用于 [Comfy Platform](https://platform.comfy.org/profile/api-keys)）
+
+### RunningHub 云端配置
+
+- `runninghub_api_key`: RunningHub API 密钥（使用云端工作流时必填）
+- `runninghub_concurrent_limit`: 并发执行限制（1-10，普通会员默认为 1）
+- `runninghub_instance_type`: 实例类型（可选）
+  - 留空或不设置：使用 24GB 显存机器
+  - `"plus"`: 使用 48GB 显存机器（适合大尺寸视频生成）
 
 ### 图像配置
 
@@ -65,6 +77,12 @@ comfyui:
 
 ---
 
+## 模板配置
+
+- `default_template`: 默认帧模板路径（例如 `1080x1920/image_default.html`）
+
+---
+
 ## 更多信息
 
 配置文件会自动在首次运行时创建。
diff --git a/docs/zh/user-guide/web-ui.md b/docs/zh/user-guide/web-ui.md
index a0cad90..a23ad9d 100644
--- a/docs/zh/user-guide/web-ui.md
+++ b/docs/zh/user-guide/web-ui.md
@@ -11,6 +11,7 @@ Web 界面采用三栏布局：
 - **左侧栏**: 内容输入与音频设置
 - **中间栏**: 语音与视觉设置  
 - **右侧栏**: 视频生成与预览
+- **侧边栏**: 系统配置与 FAQ
 
 ---
 
@@ -27,6 +28,14 @@ Web 界面采用三栏布局：
 - **AI 生成内容**: 输入主题，AI 自动创作文案
 - **固定文案内容**: 直接输入完整文案
 
+### 固定文案分割方式
+
+使用固定文案模式时，可选择文案的分割方式：
+
+- **按段落分割**: 以空行为分隔符，每个段落作为一个分镜
+- **按行分割**: 以换行符为分隔符，每行作为一个分镜
+- **按句子分割**: 智能识别句子边界，每句话作为一个分镜
+
 ### 背景音乐
 
 - 支持内置音乐
@@ -50,17 +59,19 @@ Web 界面采用三栏布局：
 
 ## 视觉设置
 
-### 图像生成
+### 图像/视频生成
 
-- 选择图像生成工作流
-- 设置图像尺寸
+- 选择媒体生成工作流（图像或视频）
 - 调整提示词前缀控制风格
 
 ### 视频模板
 
-- 选择视频模板
-- 支持竖屏/横屏/方形
-- 可预览模板效果
+- **模板预览画廊**: 可视化预览所有可用模板
+- 支持竖屏 (1080x1920) / 横屏 (1920x1080) / 方形 (1080x1080)
+- 模板类型：
+  - `static_*.html`: 静态模板（无 AI 生成媒体）
+  - `image_*.html`: 图像模板（需要 AI 生成图像）
+  - `video_*.html`: 视频模板（需要 AI 生成视频）
 
 ---
 
@@ -69,9 +80,19 @@ Web 界面采用三栏布局：
 点击「生成视频」按钮后，系统会：
 
 1. 生成视频文案
-2. 为每个分镜生成配图
+2. 为每个分镜生成配图/视频
 3. 合成语音解说
 4. 合成最终视频
 
 生成完成后自动预览。
 
+---
+
+## FAQ
+
+侧边栏内置了常见问题解答（FAQ），点击可快速查看：
+
+- 常见配置问题
+- 生成失败解决方案
+- 性能优化建议
+