Summary

This project provides an intelligent Agent system for Xiaohongshu (小红书) content operations, automating the entire workflow from data acquisition to content generation and publishing. It targets content operators, self-media automation practitioners, and AI Agent developers, offering a streamlined and intelligent approach to content production by combining data analysis with generative AI capabilities.

How It Works

The system integrates data analysis with large language model generation for automated content production. It utilizes Playwright for web scraping Xiaohongshu content, filtering for text and image posts while avoiding videos and ads. Analyzed data yields insights into high-frequency keywords, popular tags, title structures, and user preferences. These insights inform LLM-driven topic generation, followed by the creation of complete content drafts including titles, body text, tags, interaction prompts, and image suggestions. Image generation is handled by OpenAI's gpt-image-1, and content can be published via MCP protocol or REST API, forming a complete, intelligent content pipeline.

Quick Start & Requirements

Installation involves setting up a Python virtual environment, installing dependencies (pip install -r requirements.txt), and installing Playwright browsers (playwright install chromium). Environment variables (.env) require at least an OpenAI API key and model configuration. The service is started with uvicorn app.main:app --reload --host 127.0.0.1 --port 8000, accessible via Web UI at http://127.0.0.1:8000 and API documentation at http://127.0.0.1:8000/docs. Prerequisites include Python, an OpenAI API key with access to gpt-4o-mini and gpt-image-1, Playwright, local Xiaohongshu login cookies (data/raw/xhs_cookies.json), and a running local Xiaohongshu MCP service for publishing.

Highlighted Details

• End-to-End Automation: Manages data crawling, trend analysis, topic suggestion, text and image content generation, and direct publishing.
• MCP Server Integration: Encapsulates the content pipeline as an MCP tool, enabling integration with AI assistants like Claude Desktop and Cursor.
• Web UI: Offers a graphical interface for managing the crawling, generation, and publishing workflow.
• Feishu Synchronization: Optional feature to sync crawled data and generated content to Feishu tables for team collaboration and review.
• Data Analysis Module: Extracts structured insights including keywords, tags, title patterns, and user observations from scraped content.

Maintenance & Community

No specific details regarding maintainers, community channels (e.g., Discord, Slack), or roadmap are provided.

Licensing & Compatibility

License type and compatibility notes for commercial use are not specified.

Limitations & Caveats

Crawling functionality necessitates pre-existing Xiaohongshu login cookies saved locally. Publishing requires a locally running Xiaohongshu MCP service and successful QR code login. Image generation depends on a valid OpenAI API key with appropriate model access. The Feishu synchronization feature is optional and requires correct configuration.