xiaohongshu

xiaohongshu-mcp is an MCP server that automates interactions with Xiaohongshu (Little Red Book) through a modular set of capabilities. It supports login status checks, posting and publishing 图文 and 视频 content (with options for local file paths or image/video inputs), content search, fetching home page recommendations, retrieving full post details (including interactions and comments), posting comments, liking/unliking, and bookmarking/favoriting notes. The tooling is designed to work via MCP interfaces or HTTP API calls, enabling automated workflows for content creation, engagement, and data retrieval on Xiaohongshu.

To use the server, deploy it via Docker (the recommended approach for this MCP), then interact with its endpoints or the provided MCP tools to perform operations such as logging in, publishing posts, and moderating engagement. For example, you can publish a 图文 post by supplying a title, description, and image list (either HTTP URLs or local file paths), or publish a 本地视频 by providing the absolute path to the video file. You can also fetch the current feed or search results, inspect post details including comments and likes, and perform actions like commenting, liking, or bookmarking posts. The server exposes both HTTP API and MCP tool interfaces to accommodate various automation flows.

If you’re automating workflows, consider common sequences such as: ensure login first, fetch a set of target posts or a hashtag feed, create engaging posts with appropriate tags, monitor engagement and respond with comments or likes as needed, and finally gather performance metrics from post and user data.

Prerequisites:

• Docker installed on your host (recommended) or alternative runtime if you prefer a different deployment method.
• Basic knowledge of using Docker commands.

Installation steps (Docker, recommended):

1. Pull and run the MCP image: docker pull xpzouying/xiaohongshu-mcp docker run -it xpzouying/xiaohongshu-mcp

2. If you want to use Docker Compose (optional but convenient), create a docker-compose.yml with: version: '3' services: xiaohongshu-mcp: image: xpzouying/xiaohongshu-mcp container_name: xiaohongshu-mcp stdin_open: true tty: true environment: - LOG_LEVEL=info

3. Prerequisites for local builds (optional):

   • Go (for compiling from source)
   • Node.js, Python, or other tools only if you plan to use alternative deployment methods mentioned in the README.

4. Access the MCP server:

   • If using Docker run: follow the container’s stdout to see how to connect, or map ports as needed when starting the container.

5. Alternative installation methods described in the repository wrapper (if you prefer not to use Docker):

   • Build from source with Go (refer to the README’s sections for GOPROXY setup and go build steps)
   • Use prebuilt binaries for macOS/Windows/Linux as provided in the repository releases.

Tips and common issues:

• Always ensure you are logged in before performing actions that require authentication, such as posting content, commenting, or liking.
• When publishing 图文, keep in mind Xiaohongshu’s constraints: titles should be under 20 Chinese characters and the body under 1000 characters.
• For video posting, only local absolute paths are supported by the MCP tooling; remote image/video URLs may be used for image inputs but video uploads typically require local files.
• If you encounter login expiration or cookies issues, re-authenticate using the provided login workflow and then retry actions.
• When using the Docker deployment, remember to expose any needed ports and consider volume mounting for persistent data if your workflow requires it.
• The MCP supports two Openclaw-enabled paths: xiaohongshu-mcp-skills (for deployed users) and xiaohongshu-skills (out-of-the-box). Choose the path that matches your deployment scenario.
• Review the DONATIONS and CONTRIBUTING documentation linked in the repository for maintenance and contribution guidelines.