// the find
JoeanAmier/XHS-Downloader
小红书（XiaoHongShu、RedNote）链接提取/作品采集工具：提取账号发布、收藏、点赞、专辑作品链接；提取搜索结果作品、用户链接；采集小红书作品信息；提取小红书作品下载地址；下载小红书作品文件
XHS-Downloader is a Python tool for scraping and downloading content from Xiaohongshu (RedNote/XiaoHongShu), a Chinese social platform. It supports bulk extraction of posts, images, and videos from user profiles, search results, and collections — with a TUI, REST API, and MCP server all baked in. Squarely aimed at Chinese users or developers who need to archive or process XHS content programmatically.
Multiple consumption modes (TUI, FastAPI, MCP, CLI, browser userscript) from a single codebase is genuinely well-executed for a scraping tool. The author-archive feature with automatic nickname updates when a creator renames themselves is a thoughtful detail that most download tools ignore. Using uv for dependency management shows current Python packaging hygiene. The download-record deduplication via SQLite means re-running it on the same account won't re-download gigabytes of existing files.
The cookie handling is manual and already partially broken — browser auto-read was deprecated and stripped out, so you're stuck copying cookies from DevTools every time your session expires, which is a maintenance tax for any automation use case. The tool reverse-engineers a private API with no official support, meaning any platform update can silently break it, and there's no abstraction layer to isolate that breakage from the rest of the code. The internationalization support (zh_CN / en_US) is surface-level — documentation, issues, and community are almost entirely Chinese, making it practically inaccessible for non-Chinese contributors. Resume-on-failure for downloads exists, but the retry logic caps at 5 attempts with no exponential backoff, which will hammer a rate-limited endpoint.