xhs-web-note-capture

What problem does it solve?

Capture complete Xiaohongshu (小红书) note content that is only available from a real logged-in browser session, including hidden or image-heavy posts, and convert them into local markdown files with metadata and OCRed text so the full post can be archived, searched, or edited offline.

Core Features & Use Cases

• Logged-in capture: Navigate from an authenticated author or search page to obtain durable pc_user URLs containing xsec_token and save the full HTML with cookies.
• State parsing: Extract structured note data by parsing window.INITIAL_STATE to reliably read title, body, tags, imageList, and interactInfo.
• Image handling and OCR: Download images from imageList, run OCR (supports a bundled macOS Vision OCR binary), and include raw OCR output alongside original image links.
• Output: Produce one markdown file per post containing metadata, full正文, image links, and OCR text; useful for archiving private notes, research, and content migration.

Quick Start

Save the logged-in note page HTML with cookies, then run the extractor to convert the saved HTML into a markdown file in your chosen output directory.