ShortVideo.AutoPublisher

The system uses an interface-based architecture to handle different social media platforms. All platform-specific agents implement the IAiAgent interface.

IAiAgent Interface Methods

• LoginAsync(): Performs login verification.
• UploadVideoAsync(): Handles video file uploading.
• SetVideoMetadataAsync(): Sets title, description, and tags.
• SetCoverImageAsync(): Sets the video cover image.
• PublishAsync(): Finalizes the publishing process.

Architecture Layers

1. IAiAgent Interface: Defines the standard contract for all agents.
2. AgentBase Class: Provides common functionality including:
   • Retry strategies (via Polly)
   • Browser control (via Playwright)
   • Captcha detection and handling
   • Unified status and progress events
3. Platform-Specific Implementations: Concrete classes like DouyinAgent, XiaohongshuAgent, BaijiahaoAgent, WeixinChannelAgent, and ToutiaoAgent located in src/ShortVideo.AutoPublisher.Core/OpenClaw/Agents/.

Supported Platforms

• Douyin ✅
• Xiaohongshu ✅
• Baijiahao ✅
• Weixin Channel ✅
• Toutiao ✅

The project follows a layered architecture built on .NET 8.0 and WPF-UI 3.0.4 using the MVVM pattern.

Architecture Layers:

• Views: WPF-UI 3.0.4 with Fluent Design.
• ViewModels: CommunityToolkit.Mvvm for data binding.
• Services: Core logic including VideoContentService and PublishService.
• OpenClaw AI Agent Module: Handles IAiAgent implementations, platform agents, task scheduling, and browser management via Microsoft.Playwright.
• Infrastructure: Data persistence via SQLite/Dapper, network monitoring, and file downloading.
• Domain: Core entities, enums, and repository interfaces.

Key Dependencies:

• Browser Automation: Microsoft.Playwright (v1.40.0)
• Data Storage: SQLite (v1.0.118) and Dapper (v2.1.35)
• Resilience: Polly (v8.2.0) for transient fault handling
• Image Processing: SixLabors.ImageSharp (v3.1.6) for cover cropping
• Logging: Serilog (v3.1.1)

To authenticate, you must provide Cookie data for the target platform.

Method 1: Browser Developer Tools

1. Log in to the platform's creator center (e.g., https://creator.douyin.com).
2. Press F12 to open Developer Tools.
3. Go to the Network tab and refresh the page.
4. Select any request and find the Cookie field in the Request Headers. Copy the full value.

Method 2: Browser Extensions Use extensions like EditThisCookie or Cookie Editor.

1. Log in to the platform.
2. Open the extension and select Export (choose JSON format if required, though the app often accepts the raw string).
3. Paste the exported data into the application.

Platform Creator Center URLs:

• Douyin: https://creator.douyin.com
• Xiaohongshu: https://creator.xiaohongshu.com
• Baijiahao: https://baijiahao.baidu.com
• Weixin Channel: https://channels.weixin.qq.com
• Toutiao: https://mp.toutiao.com

To use the automated publishing features, you must provide valid Cookies for each platform.

Method 1: Browser Developer Tools

1. Open your browser and log in to the target platform's creator center (e.g., Douyin Creator Center).
2. Press F12 to open Developer Tools.
3. Switch to the Network tab.
4. Refresh the page and select any request.
5. In the Request Headers, find the Cookie field and copy the full value.

Method 2: Browser Extensions (Recommended)

Use extensions like EditThisCookie or Cookie Editor (Chrome) or Cookie Quick Manager (Firefox).

1. Install the extension and log in to the target platform.
2. Click the extension icon.
3. Select Export and choose the JSON format.
4. Paste the exported data into the application.

Target Creator Center URLs

• Douyin: https://creator.douyin.com
• Xiaohongshu: https://creator.xiaohongshu.com
• Baijiahao: https://baijiahao.baidu.com
• Weixin Channel: https://channels.weixin.qq.com
• Toutiao: https://mp.toutiao.com

You can expose the AutoPublisher functionality to AI Agents (like Claude Desktop) using the Model Context Protocol (MCP).

1. Configure Claude Desktop

Edit your claude_desktop_config.json file:

• Windows: %APPDATA%\Claude\claude_desktop_config.json
• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

Add the following configuration (ensure the path to your project is correct):

{
  "mcpServers": {
    "autopublisher": {
      "command": "dotnet",
      "args": ["run", "--project", "D:/Git/ShortVideo.AutoPublisher/src/ShortVideo.AutoPublisher.Mcp"]
    }
  }
}

2. Available MCP Tools

Once connected, the AI Agent can call the following tools:

• list_videos: Get video library list
• add_video: Add video to library
• get_video: Get video details
• delete_video: Delete video
• search_videos: Search videos
• list_accounts: Get platform account list
• add_account: Add platform account
• delete_account: Delete account
• validate_account: Validate account status
• publish_video: Publish video to a specific platform
• list_tasks: Get publishing task list
• get_task_status: Query task status
• retry_task: Retry failed task
• cancel_task: Cancel task
• scheduler_start: Start task scheduler
• scheduler_stop: Stop task scheduler

The project includes a Model Context Protocol (MCP) server that allows AI agents like Claude Desktop to control the publisher directly.

Setup for Claude Desktop: Add the following to your claude_desktop_config.json:

Windows: %APPDATA%\Claude\claude_desktop_config.json
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "autopublisher": {
      "command": "dotnet",
      "args": ["run", "--project", "D:/Git/ShortVideo.AutoPublisher/src/ShortVideo.AutoPublisher.Mcp"]
    }
  }
}

Available MCP Tools:

• list_videos, add_video, get_video, delete_video, search_videos (Video Library)
• list_accounts, add_account, delete_account, validate_account, set_default_account (Accounts)
• publish_video (Publishing)
• list_tasks, get_task_status, retry_task, cancel_task (Tasks)
• scheduler_start, scheduler_stop (Scheduler)