aicut-editing

name	aicut-editing
description	指导如何进行项目管理（新建、查看、切换、删除）、操作 AIcut 时间轴、添加媒体、创建字幕及导出视频。当用户询问关于项目列表管理、新建/切换/删除项目、视频剪辑、时间轴操作、添加素材、生成的视频有问题或媒体管理时使用此技能。

AIcut Video Editing Skill

This skill helps understand and manipulate the AIcut video editing system.

Core Concepts

Project Structure

• Project Snapshot: ai_workspace/project-snapshot.json - The single source of truth for timeline state.
• Project Folders: projects/<display-name>/ - Readable folder names mapped to internal IDs via projects/projectIdMap.json.
• Assets: projects/<display-name>/assets/ - Media files organized by type.
• Exports: exports/ - Final rendered videos.

Project Management

1. UI Creation & Management

• Create: Click "New Project" in the /projects page.
• View: Browse existing projects in the UI list. The UI synchronizes with the filesystem in real-time.
• Delete: Click the Delete icon on a project card. This removes the project from the UI and deletes its folder from the projects/ directory.

2. Programmatic Lifecycle (New/Switch/Delete Project Logic)

To manage projects via API/Python:

1. Initialize: Call switch_project(projectId). If the project is loaded for the first time, it creates the folder. If it exists, the UI will automatically jump to that project.
2. Archiving (Readable Folders): Use archive_project(). The system uses the project's name as the folder name.
3. Switching: Call switch_project(anotherId). The UI will perform a hot-reload and navigate to the new project editor.
4. Mapping: Folder names are mapped to internal IDs in projectIdMap.json.
5. Deletion: Call delete_project(projectId) to permanently remove the project folder and its ID mapping.

Element Types

1. Media Element (Video/Image/Audio)

{
  "id": "e.g., el-0",
  "type": "media",
  "mediaId": "asset-reference-id",
  "name": "display-name.mp4",
  "thumbnailUrl": "/api/media/serve?path=...", // CRITICAL for video/image preview
  "startTime": 0.0,      // In seconds
  "duration": 5.0,       // Visible duration
  "trimStart": 0.0,      // Crop from start of source
  "trimEnd": 0.0,        // Crop from end of source
  "x": 960, "y": 540,    // Postion (Center origin is 960, 540)
  "scale": 1.0, 
  "opacity" 
   
   
      
  

2. Text/Subtitle Element

To ensure compatibility between Python and Frontend, use this format:

{
  "id": "unique-id",
  "type": "text",
  "content": "The display text",    // Primary text field
  "startTime": 0.0,
  "duration": 3.0,
  "trimStart": 0, "trimEnd": 0,    // MUST include (default to 0)
  "x": 960, "y": 900,              // Subtitle position (bottom center)
  "rotation": 0, "opacity": 1,     // MUST include defaults
  "fontSize": 60,                  // Styles MUST be at top level
  "fontFamily": "Arial"           
   
   
   
                          
      
      
  

Data Schema Rules (The "Contract")

1. Primary Text Field: Use content as the single source of truth for text. The name field is OMITTED for text elements to prevent redundancy.
2. Flattened Styles: While Python SDK uses a style object, the frontend requires fontSize, color, etc., at the top level of the element.
3. Implicit Defaults: Never skip trimStart, trimEnd, rotation, and opacity. If missing, the frontend might fail to render or default to 0 (invisible).
4. Coordinate System: origin (0,0) is Top-Left. 1920x1080 canvas. Standard center is (960, 540).
5. IDs: IDs must be unique strings (e.g., nanoid or sub-0, sub-1).

Common Operations

Adding Media

Preferred Method: Use AIcut SDK (import_media). The SDK automates the following steps:

1. Thumbnail Generation: Calls /api/media/generate-thumbnail for videos to create high-quality frames.
2. Asset Creation: Automatically calculates IDs and creates an asset entry with a stable serve URL.
3. Element Placement: Adds the element to the timeline with the correct thumbnailUrl.

Manual Method (Advanced):

1. Create an asset entry in the assets array.
2. Create an element in a track referencing the asset's id.
3. MUST set thumbnailUrl (use the asset's URL or a dedicated JPG thumb) for visual feedback.

Creating Subtitles

1. Create or use an existing text track (type: text).
2. Add text elements according to the Text Element Schema above.
3. Synchronize startTime with audio assets.

Modifying Volume

• Set element.volume (0.0 = mute, 1.0 = normal, up to 10.0 = 1000%).

API Integration

The frontend exposes a state-syncing API at POST /api/ai-edit. Always prefer using the Python SDK (scripts/aicut_sdk.py) over raw HTTP requests.

Supported Actions (POST)

Action	Description	Key Data Fields
addSubtitle	Add a single subtitle	text, startTime, duration, fontSize, color
addMultipleSubtitles	Batch add subtitles	subtitles (Array of subtitle objects)
clearSubtitles	Remove subtitles	startTime, duration (Optional range)
switchProject	Load or create project	projectId
archiveProject	Save workspace to disk	projectId (Optional)
deleteProject	Delete project folder	projectId
updateSnapshot	Push full timeline state	project, tracks, assets (Full JSON object)

Media Utilities

• GET /api/media/serve?path=<encoded_absolute_path>: Serves local files as browser-accessible URLs.
• POST /api/media/generate-thumbnail: Generate a JPG from a video file (Body: { filePath: string }).

Snapshot & Project Endpoints

• GET /api/ai-edit?action=getSnapshot: Retrieve current timeline state.
• GET /api/ai-edit?action=listProjects: List all archived projects.
• GET /api/ai-edit?action=poll: Polled by frontend to fetch pending edits.

Best Practices

1. Avoid curl on Windows: Use a temporary Python script or the AIcut SDK for stable JSON communication.
2. Paths: DO NOT use /materials/ relative paths. Always use absolute paths wrapped in the /api/media/serve?path= API. This ensures files from any drive (F:, D:, etc.) work correctly.
3. Atomic Updates: Read the full snapshot, modify your slice, then write the full snapshot back.
4. Validation: Ensure all numeric fields are number type, not string.
5. Layering: Text tracks should remain at the top (lower array index) to be visible over media.
6. SDK First: Always use scripts/aicut_sdk.py. It correctly handles thumbnail generation calls and internal path mapping.

Executable Scripts

The following scripts/CLIs are available in the scripts/ directory:

Tool	Purpose	Usage Example
aicut_tool.py	Centralized CLI for all common tasks.	python scripts/aicut_tool.py media import "path/to/video.mp4"
aicut_sdk.py	Core library for all programmatic edits.	from aicut_sdk import AIcutClient
fix_subtitle_pos.py	Align subtitles to bottom center.	python scripts/fix_subtitle_pos.py
project_manager.py	Standalone project lifecycle tool (legacy support)	python scripts/project_manager.py list

Note: These scripts depend on scripts/aicut_sdk.py. Ensure you run them from the skill directory or add the directory to your PYTHONPATH.