Xiaohongshu MCP: The Complete Guide to AI-Powered Content Management for Little Red Book

Xiaohongshu (小红书, also known as "Little Red Book" or RED) is China's dominant lifestyle and social commerce platform with over 300 million monthly active users. For content creators and marketers operating on this platform, managing posts, engagement, and analytics can be overwhelming. Xiaohongshu MCP solves this by providing a Model Context Protocol (MCP) server that enables AI assistants like Claude, ChatGPT, and Cursor to directly interact with your Xiaohongshu account — publishing content, searching posts, managing comments, and tracking engagement, all through natural language commands.

With 11,200+ GitHub stars, 1,678 forks, 90 releases, and a thriving community, this project by xpzouying has become the most popular social media MCP server on GitHub. Built in Go and powered by browser automation, it operates through the user's own authenticated session — no unofficial API hacking involved.

---

What Is Xiaohongshu MCP?

Xiaohongshu MCP is an open-source MCP server written in Go that bridges AI assistants with the Xiaohongshu platform. Using the standardized Model Context Protocol — often described as the "USB port for AI tools" — it exposes a set of tools that any MCP-compatible AI client can call to perform actions on your Xiaohongshu account.

The server uses Playwright-based browser automation to interact with the official Xiaohongshu web interface through the user's own login session (via QR code authentication). This means it operates within the platform's normal usage patterns — no reverse engineering, no unofficial APIs.

Key Stats

Metric	Value
GitHub Stars	11,200+
Forks	1,678
Releases	90
Language	Go
Created	August 2025
Last Updated	March 2026 (daily)
Homepage	haha.ai/xiaohongshu-mcp
Topics	mcp, mcp-server, xiaohongshu-mcp
Docker	xpzouying/xiaohongshu-mcp

---

Core Features

1. Content Publishing

Publish image, text, and video content directly to Xiaohongshu through your AI assistant.

Image + Text Posts (publish_content):

• Title (required, max 20 characters per XHS rules)
• Content (required, max 1,000 characters)
• Images (required, at least 1) — supports HTTP URLs and local file paths (local recommended)
• Tags (optional) — adds hashtags for discoverability
• Schedule (optional) — ISO8601 format, 1 hour to 14 days ahead
• Visibility — public, self-only, or mutual-follow friends only
• Products — product keyword tagging for e-commerce integration
• Original declaration (optional)

Video Posts (publish_with_video):

• Local video file path (MP4, max ~1GB recommended)
• Same metadata options as image posts
• Auto-waits for video processing before publishing

2. Content Discovery

• search_feeds — Keyword search with advanced filters:
  • Sort by: relevance, newest, most liked, most commented, most bookmarked
  • Note type: all, video only, image+text only
  • Time range: all, 24h, 1 week, 6 months
  • Scope: all, viewed, unviewed, followed accounts
  • Location: all, same city, nearby
• list_feeds — Get personalized homepage recommendations
• get_feed_detail — Full post details with engagement data, comments, and sub-replies

3. Engagement Management

• post_comment_to_feed — Auto-comment on posts
• reply_comment_in_feed — Reply to specific comments (by comment_id or user_id)
• like_feed / favorite_feed — Like/unlike, bookmark/unbookmark with smart duplicate detection

4. User Intelligence

• user_profile — Retrieve user profiles including:
  • Basic info (nickname, bio, avatar, verification)
  • Stats (followers, following, likes, post count)
  • Published content list

5. Session Management

• check_login_status — Verify current login state
• get_login_qrcode — Get login QR code (returns Base64 image)
• delete_cookies — Reset session for re-login

---

Getting Started

Option A: Pre-compiled Binaries (Simplest)

Download from GitHub Releases — available for macOS (ARM/Intel), Windows, and Linux.

# 1. Run login tool (first time)
chmod +x xiaohongshu-login-darwin-arm64
./xiaohongshu-login-darwin-arm64

# 2. Start MCP server
chmod +x xiaohongshu-mcp-darwin-arm64
./xiaohongshu-mcp-darwin-arm64

Option B: Docker (Recommended for Production)

# Pull from Docker Hub
docker pull xpzouying/xiaohongshu-mcp

# Or use docker-compose
wget https://raw.githubusercontent.com/xpzouying/xiaohongshu-mcp/main/docker/docker-compose.yml
docker compose up -d

Docker auto-configures Chrome, Chinese fonts, cookie storage, and exposes port 18060.

Option C: Build from Source

git clone https://github.com/xpzouying/xiaohongshu-mcp.git
cd xiaohongshu-mcp

# Build and run
go run cmd/login/main.go   # Login first
go run .                    # Start MCP server

Option D: Browser Extension (Zero Config)

For non-technical users: x-mcp — a browser extension (Chrome/Edge) that works without Docker or server setup.

---

MCP Client Integration

Supported Clients

The server works with any MCP-compatible client:

Client	Status
Claude Desktop	✅ Fully supported
Claude Code	✅ Fully supported
Cursor	✅ Fully supported
Windsurf	✅ Supported
Cherry Studio	✅ Supported
Trae	✅ Supported
n8n	✅ Supported
Spring AI	✅ Supported
LangChain	✅ Supported

Configuration Example

{
  "mcpServers": {
    "xiaohongshu-mcp": {
      "command": "./xiaohongshu-mcp-darwin-arm64",
      "args": []
    }
  }
}

For remote/Docker setups, use SSE transport:

{
  "mcpServers": {
    "xiaohongshu-mcp": {
      "url": "http://localhost:18060/sse"
    }
  }
}

---

Ecosystem & Integrations

OpenClaw Integration

The project offers deep integration with OpenClaw (the AI agent framework):

• xiaohongshu-mcp-skills — For users who already have xiaohongshu-mcp deployed
• xiaohongshu-skills — Out-of-the-box version requiring no prior setup

Community & Support

• Feishu (Lark) group for mutual help
• WeChat groups for community discussion
• Active GitHub Discussions
• Contributing Guide

---

Real-World Results

The author shares verified results from production use:

• Day 1: 999+ likes and bookmarks on published content
• 1 week: Significant engagement growth with AI-automated content operations
• Stable operation: Over 1 year of stable running on the author's own account, with no bans — only occasional cookie expirations requiring re-login
• The author uses Claude Code for fully automated content operations

Operational Tips from the Author

1. Titles must be ≤ 20 characters — Xiaohongshu enforces this strictly
2. Content body ≤ 1,000 characters
3. Image+text posts outperform video and pure text for recommendations
4. Tags increase discoverability — use relevant hashtags
5. Daily post limit: ~50 posts per day per account
6. Single session only: Don't log into the web version elsewhere while MCP is running
7. Avoid prohibited words — use third-party tools to check before posting
8. Never do引流 (traffic diversion) or pure reposts — Xiaohongshu actively penalizes this

---

Security & Privacy

1. Local credentials: Cookies and Bearer tokens stored locally in auth.json — never committed to Git
2. Session-based: Uses the user's own authenticated web session
3. No unofficial APIs: Interacts through the official web interface
4. Rate limit aware: Web APIs have natural rate limits; not suitable for high-frequency bulk operations
5. Proxy support: XHS_PROXY environment variable for HTTP/HTTPS/SOCKS5 proxies

---

Xiaohongshu MCP vs Alternatives

This is an MCP server for social media automation. Here's how it compares:

Feature	Xiaohongshu MCP	OpenTweet MCP	Ayrshare MCP	Postiz MCP	Xpoz MCP
Stars	11.2K	1K+	N/A (SaaS)	2K+	N/A (SaaS)
Platform	Xiaohongshu	Twitter/X	13+ platforms	Multi-platform	Multi-platform
Publishing	✅ Image+Video	✅ Tweets	✅ All formats	✅ Scheduling	❌ Read-only
Search	✅ Advanced filters	✅ Basic	❌	❌	✅ Intelligence
Comments	✅ Post+Reply	❌	❌	❌	❌
Like/Bookmark	✅ Smart toggle	❌	❌	❌	❌
User Profiles	✅ Full data	❌	❌	❌	✅
Self-hosted	✅ Go/Docker	✅	❌ SaaS	✅	❌ SaaS
Open Source	✅	✅	❌	✅	❌
Chinese Platform	✅ Native	❌	❌	❌	❌

When to Choose Each

• Xiaohongshu MCP: The only option for Xiaohongshu automation. Unmatched engagement features (comments, likes, bookmarks) and deep Chinese platform integration.
• OpenTweet MCP: Best for Twitter/X-focused automation with scheduling and analytics.
• Ayrshare MCP: Best for enterprises needing a single SaaS to post across 13+ Western platforms.
• Postiz MCP: Best open-source alternative for multi-platform scheduling (Western platforms).
• Xpoz MCP: Best for social media intelligence and research across Twitter, Instagram, TikTok, Reddit.

---

Frequently Asked Questions

Will my account get banned?

The author reports over 1 year of stable operation with no bans. New accounts without identity verification may trigger verification prompts (standard behavior, not a ban).

Which AI client should I use?

Claude Code provides the best experience for automated operations. Claude Desktop and Cursor are also fully supported.

How do I handle cookie expiration?

Sessions expire periodically. Use delete_cookies to clear the session, then re-login via QR code.

Can I use this in production?

Yes. The author runs it in production. Use Docker for the most stable deployment. Respect Xiaohongshu's daily post limits (~50/day).

---

Conclusion

Xiaohongshu MCP is the most comprehensive social media MCP server available on GitHub. With 11,200+ stars, 90 releases, and a feature set that covers the full lifecycle of content operations — publishing, discovery, engagement, and user intelligence — it sets the standard for what AI-powered social media management can look like.

For anyone operating on Xiaohongshu — whether managing a personal brand, running an e-commerce storefront, or producing content at scale — this tool transforms your AI assistant into a full-featured social media operations platform.

---

Resources

• Xiaohongshu MCP on GitHub
• Author's Blog
• Docker Hub Image
• x-mcp Browser Extension
• OpenClaw Skills
• Contributing Guide