Dy Xhs MCP Server

🚀 Douyin and Xiaohongshu Content Extraction MCP Server

A server based on the Model Context Protocol (MCP) that supports extracting video, image, and text content from Douyin and Xiaohongshu sharing links.

📋 Project Declaration

Official documentation address: https://github.com/yzfly/douyin-mcp-server

Please refer to the README.md file of this project for detailed information on the project's functional features, usage methods, API configuration instructions, etc.

Important reminder: If a third - party platform cannot be used normally due to limitations in its support for the MCP Server function, please contact the corresponding platform. This project does not provide any form of technical support or guarantee, and users are required to bear any losses or damages that may arise from using this project on their own.

Legal statement:

1. This project is released under the Apache 2.0 license.
2. This project is for learning and research purposes only and shall not be used for any illegal or违规 purposes.
3. The use of this project must comply with relevant laws and regulations.
4. The author and contributors of this project shall not be liable for any part of the project.

✨ Features

Douyin Features

• 🎵 Watermark - free video acquisition - Obtain high - quality watermark - free videos from Douyin sharing links.
• 🎧 Intelligent audio extraction - Automatically extract audio content from videos.
• 📝 AI text recognition - Use advanced speech recognition technology to extract text content.

Xiaohongshu Features

• 📱 Video/image - text extraction - Support Xiaohongshu video notes and image - text notes.
• 📝 Copywriting extraction - Automatically extract the copywriting content of notes.
• 🖼️ Image acquisition - Extract all image links in notes (automatically remove watermark parameters).
• 🎤 Video speech recognition - Extract speech text from video notes (API key required).
• 🎬 Watermark - free video - Automatically remove watermark parameters from video URLs to obtain high - quality watermark - free videos.
• 🤖 AI watermark removal - Use AI technology to intelligently detect and remove watermarks in videos (OpenCV required).
• 🔍 Intelligent parsing - Support multiple link formats and automatically identify note types and content.

General Features

• 🧹 Automatic cleanup - Intelligently clean up temporary files during the processing.
• 🔧 Flexible configuration - Support custom API configuration. By default, it uses Alibaba Cloud Bailian API.

🚀 Quick Start

Step 1: Get the API Key

Go to Alibaba Cloud Bailian API to get your API key:

Step 2: Configure the API Key

Create a config.json configuration file:

{
  "api_key": "sk-your-api-key-here",
  "model": "paraformer-v2",
  "language_hints": ["zh", "en"],
  "temp_dir": "temp"
}

Step 3: Configure Claude Desktop

Add the following configuration to the configuration file of applications that support MCP Server, such as Claude Desktop and Cherry Studio:

{
  "mcpServers": {
    "douyin-mcp": {
      "command": "python3.11",
      "args": ["-m", "douyin_mcp_server.server"]
    }
  }
}

Step 3: Start Using

After the configuration is completed, you can normally call the MCP tool in supported applications.

🔧 API Configuration Instructions

Current Version (>= 1.2.0)

The latest version uses the Alibaba Cloud Bailian API by default, which has the following advantages:

• ✅ Better recognition effect.
• ✅ Faster processing speed.
• ✅ Less local resource consumption.

Configuration steps:

1. Go to Alibaba Cloud Bailian to activate the API service.
2. Get the API Key and configure it in the api_key field of the config.json file.

Compatibility with Older Versions (<= 1.1.0)

If you need to use an older version, use the following configuration:

{
  "mcpServers": {
    "douyin-mcp": {
      "command": "uvx",
      "args": ["douyin-mcp-server@1.1.0"],
      "env": {
        "DOUYIN_API_KEY": "your-api-key-here"
      }
    }
  }
}

Note: The older version uses the Silicon Flow API. You need to register an account on Silicon Flow and get the API Key.

📖 Documentation for version 1.1.0

🛠️ Douyin Tool Instructions

get_douyin_download_link

Get the watermark - free download link for a Douyin video.

Parameters:

• share_link (string): Douyin sharing link or text containing the link.

Returns:

• Download link and video information in JSON format.

Features: Can be used without an API key.

extract_douyin_text

A complete text extraction tool that completes the conversion from video to text in one stop.

Processing flow:

1. Parse the Douyin sharing link.
2. Directly use the video URL for speech recognition.
3. Return the extracted text content.

Parameters:

• share_link (string): Douyin sharing link or text containing the link.
• model (string, optional): Speech recognition model. By default, paraformer-v2 is used.

Configuration requirements:

• config.json: A configuration file containing the Alibaba Cloud Bailian API key (required).

parse_douyin_video_info

A lightweight video information parsing tool.

Parameters:

• share_link (string): Douyin sharing link.

Features: Only parses basic video information without downloading the video file.

Resource Access

• douyin://video/{video_id}: Get detailed information by video ID.

🛠️ Xiaohongshu Tool Instructions

get_xiaohongshu_content

Get the complete content (video/image - text) of a Xiaohongshu note.

Parameters:

• share_link (string): Xiaohongshu sharing link or text containing the link.

Returns:

• JSON format, containing information such as video links, copywriting, images, and authors.

Features: Can be used without an API key and supports both video and image - text types.

extract_xiaohongshu_text

Extract the copywriting content of a Xiaohongshu note.

Parameters:

• share_link (string): Xiaohongshu sharing link or text containing the link.

Returns:

• The copywriting content of the note (plain text, including title, description, and tags).

Features: Can be used without an API key.

extract_xiaohongshu_video_text

Extract the speech text content from a Xiaohongshu video note.

Parameters:

• share_link (string): Xiaohongshu sharing link or text containing the link.
• model (string, optional): Speech recognition model. By default, paraformer-v2 is used.

Returns:

• The extracted text content (including copywriting and speech content).

Configuration requirements:

• config.json: A configuration file containing the Alibaba Cloud Bailian API key (required).

get_xiaohongshu_images

Get all image links in a Xiaohongshu note.

Parameters:

• share_link (string): Xiaohongshu sharing link or text containing the link.

Returns:

• A list of image links in JSON format.

Features: Can be used without an API key, and watermark parameters in image links have been automatically removed.

remove_xiaohongshu_video_watermark

Use AI technology to remove the watermark from a Xiaohongshu video.

Parameters:

• share_link (string): Xiaohongshu sharing link or text containing the link.
• output_path (string, optional): Output file path (optional, default is saved to the temp directory).

Returns:

• Processing results in JSON format, including the path of the watermark - free video, processing method, and confidence level.

Features:

• The complete AI function requires the installation of OpenCV (opencv - python).
• If OpenCV is not available, a basic processing method will be used.
• Supports intelligent detection of watermark positions and automatic repair.

Resource Access

• xiaohongshu://note/{note_id}: Get detailed information by note ID.

📦 System Requirements

Operating Environment

• Python: Version 3.10 or higher.

Dependent Libraries

• requests - HTTP request processing.
• ffmpeg-python - Audio and video processing.
• tqdm - Progress bar display.
• mcp - Model Context Protocol support.
• dashscope - Alibaba Cloud Bailian API client.
• beautifulsoup4 - HTML parsing (for Xiaohongshu content extraction).

⚠️ Notes

• 🔑 API key required: The video speech recognition function requires a valid Alibaba Cloud Bailian API key.
• 🆓 Some functions are free: Functions such as getting download links and extracting copywriting do not require an API key.
• 📱 Format support: Supports most Douyin video formats and Xiaohongshu content formats.
• 🚀 Performance optimization: Use the Alibaba Cloud Bailian API for faster and more accurate recognition results.
• 📝 Xiaohongshu support: Supports both video notes and image - text notes.

📁 Project Structure

The project files have been categorized into the following directories by type:

• docs/ - Documentation directory: Stores all project documentation (Markdown files).
• configs/ - Configuration directory: Stores all configuration files (JSON files).
• scripts/ - Script directory: Stores all executable scripts (Python and Shell scripts).
• douyin_mcp_server/ - Source code directory: The main source code of the project, using a layered architecture.

For detailed instructions, please refer to docs/DIRECTORY_STRUCTURE.md

🔧 Development Guide

Set up the Local Development Environment

# Clone the project
git clone https://github.com/AshinH/dy-xhs-mcp-server.git
cd dy-xhs-mcp-server

# Install dependencies (development mode)
pip install -e .

Run Tests

# Start the server for testing
python -m douyin_mcp_server.server

# Or use the startup script
bash scripts/start_server.sh

Local Development Configuration for Claude Desktop

Ensure that there is a config.json configuration file in the project root directory, and then add the following to the Claude Desktop configuration file:

{
  "mcpServers": {
    "douyin-mcp": {
      "command": "python3.11",
      "args": [
        "-m",
        "douyin_mcp_server.server"
      ],
      "cwd": "/path/to/your/douyin-mcp-server"
    }
  }
}

⚠️ Disclaimer

Usage Risks

• Users have full autonomy in using this project and shall bear all risks on their own.
• The author shall not be responsible for any losses, liabilities, or risks arising from users' use of this project.

Code Quality

• This project is developed based on existing knowledge and technology, and the author strives to ensure the correctness and security of the code.
• However, it is not guaranteed that the code is completely free of errors or defects, and users are required to evaluate and test it on their own.

Third - Party Dependencies

• The third - party libraries, plugins, or services that this project depends on follow their respective open - source or commercial licenses.
• Users are required to review and comply with the corresponding agreements on their own.
• The author shall not be responsible for the stability, security, and compliance of third - party components.

Legal Compliance

• Users must study relevant laws and regulations on their own to ensure that their usage behavior is legal and compliant.
• Any legal liabilities and risks arising from violations of laws and regulations shall be borne by users themselves.
• It is prohibited to use this tool to engage in any activities that infringe on intellectual property rights.
• The developer does not participate in, support, or endorse the acquisition or distribution of any illegal content.

Data Processing

• This project shall not be responsible for the compliance of users' data collection, storage, transmission, and other processing activities.
• Users shall comply with relevant laws and regulations on their own to ensure that their data processing behavior is legal and proper.

Liability Limitation

• Users shall not associate the project author, contributors, or related parties with their usage behavior.
• They shall not require the author to be responsible for any losses or damages arising from using the project.
• Secondary development, modification, or compilation programs based on this project have nothing to do with the original author.

Intellectual Property

• This project does not grant users any patent licenses.
• If using this project leads to patent disputes or infringements, users shall bear all risks and responsibilities on their own.
• Without written authorization, it shall not be used for commercial promotion, marketing, or re - authorization.

Service Termination

• The author reserves the right to terminate providing services to users who violate the statement at any time.
• Violent users may be required to destroy the obtained code and derivative works.
• The author reserves the right to update this statement without prior notice.

⚠️ Important reminder: Before using this project, please carefully read and fully understand the above disclaimer. If you have any questions or do not agree with any terms, please do not use this project. Continuing to use it means that you fully accept the above statement and voluntarily bear all risks and consequences.

📄 License

Apache License 2.0

👨‍💻 Author

• holiday - Project maintainer
• GitHub: https://github.com/AshinH

🤝 Contribution

Welcome to submit Issues and Pull Requests! We look forward to your participation and contribution.

📝 Update Log

v1.3.0 (Latest) - 2026-01-28

• 🏗️ Architecture refactoring: Adopt a layered architecture design to improve code maintainability and scalability.
  • New config/ layer: Unified configuration management.
  • New utils/ layer: Modularize utility functions (HTTP client, watermark processing).
  • New services/ layer: Separate business logic (Douyin service, Xiaohongshu service).
  • New tools/ layer: Encapsulate MCP tool functions.
  • New resources/ layer: Process MCP resources.
  • Refactor server.py: The main entry file, clearly assemble all modules.
• 📚 Documentation improvement: Add the ARCHITECTURE.md architecture description document.
• 🔧 Code optimization: Improve code reusability and testability.
• 🔚 Xiaohongshu function improvement:
  • ✨ New AI watermark removal function (remove_xiaohongshu_video_watermark).
  • 🎬 Enhance the ability to obtain watermark - free videos.
  • 🖼️ Optimize the image watermark removal algorithm.
  • 🔍 Improve the accuracy of content parsing.

v1.2.0 - 2025-XX-XX

• 🚀 Performance improvement: Faster and more accurate video copywriting extraction.
• 🔄 API upgrade: Switch to the Alibaba Cloud Bailian API, significantly improving recognition accuracy.
• 🔧 Configuration update: Update the environment variable from DOUYIN_API_KEY to DASHSCOPE_API_KEY.

v1.1.0 - 2025-XX-XX

• 🐛 Bug fix: Fix the error caused by an overly long file name when extracting videos.

v1.0.0 - 2025-XX-XX

• 🎉 Initial release: The initial version.
• ✨ Core functions: Support text extraction from Douyin videos.
• 🔗 Link acquisition: Support getting watermark - free video download links.
• 🔐 Environment configuration: Read the API key from environment variables.
• 🧹 Automatic cleanup: Automatically clean up temporary files.
• ⚙️ Flexible configuration: Support custom API configuration.