xhs-note-scraper

XHS Note Scraper

Core Rule

Use only public HTTP fetches. Do not use local browser automation, user login cookies, account sessions, or attempts to bypass access controls. If the public page does not expose SSR note data or image URLs, report that limitation and stop.

Quick Start

Run the bundled script:

node <skill-dir>/scripts/scrape_xhs_note.js "<xiaohongshu note url>" --out "<output-dir>"

Use the user's current workspace as the default output directory unless they specify another path.

The script writes:

• note.md - readable Markdown with text, metadata, and local image references
• note.json - structured note data and image URLs
• page.html - fetched public HTML
• assets/image-XX.jpg - downloaded images

Workflow

1. Confirm the URL is a Xiaohongshu note URL, typically https://www.xiaohongshu.com/explore/<note-id>....
2. Run scripts/scrape_xhs_note.js with the URL and output directory.
3. Verify the script output summary: title, author, note id, image count, and downloaded image paths.
4. If downloads fail with 403 for raw/original image attempts, keep the public CDN default image URLs exposed by the page.
5. In the final answer, link to note.md, note.json, page.html, and the assets directory. Mention any limitations such as missing comments or unavailable original-resolution files.

Notes

• SSR data is expected in window.__INITIAL_STATE__; the script extracts note.noteDetailMap[firstNoteId].note.
• Public note pages often expose text, author, tags, basic interaction counts, and images. Comments may remain dynamic and are not guaranteed.
• PowerShell can display UTF-8 Chinese incorrectly through some pipelines. Trust files written by the script as UTF-8 and verify with Node when needed.