Image-Text Hunter: Social Platform Content Scraper

Problem

When researching Xiaohongshu topics or monitoring competitors, manually searching, opening notes, copying comments, and assembling spreadsheets is slow and hard to repeat. Image-Text Hunter turns this repetitive work into executable scraping tasks: it retrieves public data using keywords, sorting rules, and note links, then produces structured logs.

How It Works

• Keyword search: set --keyword, --sort, --limit, and --output to retrieve note lists by recency, likes, comments, favorites, or other dimensions.
• Note details: pass --url to fetch a single note and use --limit to control the number of comments retrieved.
• Data output: results are saved as json or markdown logs, making them easier to reuse in analysis, reports, or scripts.

It is useful for breaking “find high-performing notes, inspect comments, and compare account strategies” into repeatable query steps.

Boundaries

• It targets public Xiaohongshu data only and does not replace logged-in account actions.
• It requires GUAIKEI_API_TOKEN, so it is not a free capability; the token should be managed carefully.
• Data is suitable for compliant market analysis, content planning, and competitor observation, not for malicious marketing, doxxing, or commercial infringement.
• It requires a Node.js environment, and logs are stored under the skill directory logs by default.