SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsxiaohongshu-lowtop

xiaohongshu-lowtop

General & other

View repository

Updated 10-15 23:57

About this skill

1. Introduction

One-sentence positioning: Helps you discover those low-follower notes on Xiaohongshu that don't have many followers but have very strong content, so you can find topics and writing styles worth learning from.

Core value：

You only need to tell me a category you're interested in, such as "home renovation" or "makeup and beauty," and I can help you pull out a batch of notes with fewer than 5,000 followers that still get high interactions. Each one comes with interaction data, publish time, and content analysis, and it also automatically summarizes the title formulas and content characteristics of this batch of viral hits, letting you quickly figure out what content is gaining traction.

Target users：

  • 📌 Xiaohongshu content creators: find topic inspiration, learn title writing, understand niche trends
  • 📌 Content operations: low-follower viral hits are the most authentic signal of user preferences, more valuable for reference than big-V data
  • 📌 MCN / brands: discover potential bloggers, assess content trends

2. Features

Core features

Feature module Capability description
📊 Category filtering Covers 25 content categories, from general trending to parenting and childcare; enter a category name or keyword to search
🏆 TOP50 list Sorted by interaction count from high to low; each note shows likes, collects, comments, and shares data
💡 Viral pattern analysis Automatically breaks down the title types and content topic distribution of this batch of notes, extracting reusable writing highlights
🔔 Daily subscription push After subscribing, the latest list of the day is automatically pushed every day at 19:30, no manual query needed
📄 File package export Generates a complete page in Xiaohongshu style, supports PDF export, convenient for saving and sharing

Data standards

Query scope: follower count below 5,000, note likes greater than 500

Each note includes: title, author, follower count, publish time, likes, comments, collects, shares, total interactions

3. Authentication

This skill calls Honghu Hub API to obtain Xiaohongshu low-follower viral data, and a valid API Key is required for normal use.

Get API Key

  1. Visit Honghu Hub official website to learn about service details
  2. Go to registration page to register an account
  3. Newly registered users will receive free credits, and can immediately start using the API service
  4. After registering and logging in, get the API Key in the personal center, in the format ak_xxxxxxxx

Configure API Key

When the skill runs, it automatically obtains the API Key through the following three-level fallback:

Priority Source Behavior
1 Current environment variable REDFOX_API_KEY Use directly
2 Shell configuration file (automatically scans ~/.zshrc、~/.bashrc, PowerShell profile, etc.) Read and inject into current environment variables
3 Not found Print detailed acquisition and configuration guidance

If not configured, please set it according to the corresponding system:

macOS / Linux：

# 将以下行追加到 ~/.zshrc（zsh）或 ~/.bashrc（bash）
export REDFOX_API_KEY=<你的apikey>
# 然后执行使其生效
source ~/.zshrc  # 或 source ~/.bashrc

Windows：

# PowerShell（管理员模式）
[Environment]::SetEnvironmentVariable("REDFOX_API_KEY", "<你的apikey>", "User")

Restart the terminal after configuration for it to take effect.

Verify configuration：

  • macOS / Linux：echo $REDFOX_API_KEY
  • Windows CMD：echo %REDFOX_API_KEY%
  • Windows PowerShell：echo $env:REDFOX_API_KEY

If you don't know how to configure it, tell the Agent to proactively help with the setup.

4. Usage Guide

4.1 Basic Queries

  1. After triggering the skill, tell me the category and date you want to check
  2. If no date is specified, the system automatically uses the most recent available data date
  3. Wait for the results to be output, including the ranking table and pattern analysis

4.2 Advanced Operations

Subscribe to daily push notifications

Reply "subscribe" or select subscribe on the results page, and you will automatically receive the latest ranking every day at 19:30.

Export file package

Reply "2" to generate a complete page file package, including all note data and layout, which can be exported as PDF.

4.3 Core Execution Flow

  1. After triggering the skill, first ask the user which category and which day's data they want to check
  2. Call the main script to get data; the script requests the API via native socket+SSL, automatically matching the category and handling the date
  3. The script obtains REDFOX_API_KEY for authentication through a three-level fallback (environment variable → Shell configuration → prompt the user)
  4. The script generates an md report file and a data cache JSON file (*_cache.json), reads all the content of the md and outputs it to the user as-is
  5. The output includes three parts: the viral note table, viral pattern analysis, and the feature selection entry point
  6. Wait for the user to choose: subscribe, export file package (reply "2"), view more, or take no action
  7. When the user replies "2": generate HTML from the cache → preview_url automatically displays it (if it does not display properly, use open_result_view as a fallback) → deliver_attachments sends the file → inform the user of the save path

4.4 Energy-saving mechanism: HTML generation does not repeatedly request the API + automatic display

When the user replies "2" to generate the HTML file package, the Agent performs the following steps:

Step 1: Generate HTML (without requesting the API)

Render directly from the already obtained data cache file without requesting the API again:

  • While generating the md report, the main script automatically outputs the *_cache.json cache file, storing the complete articles list
  • HTML generation script generate_html_on_demand.py supports the --from-cache parameter; pointing it to the cache file allows direct rendering
  • Retain the --rank_date + --keyword fallback mode, and manually request only when the cache file is unavailable

Call example：

# 节能模式：从缓存生成 HTML（0 次 API 请求）
python generate_html_on_demand.py --from-cache ./小红书居家装修低粉爆款数据_xxx_cache.json --output ./output.html

# 回退模式：重新请求 API 获取数据后渲染（仅缓存缺失时使用）
python generate_html_on_demand.py --rank_date 2026-05-26 --keyword 居家装修 --output ./output.html

Step 2: Automatically display + send HTML (must be executed; both steps are required)

After generating the HTML, the Agent must perform the following operations in order:

  1. Prefer using the preview_url tool to open the file so the user can directly see the visual page
  2. If preview_url does not display properly (including compatibility issues such as Chinese paths), immediately switch to open_result_view for fallback display
  3. Use the deliver_attachments tool to send the HTML file to the user as an attachment, ensuring the user can receive the file on different clients

None of the above operations can be omitted. After completion, inform the user of the file save path.

5. Usage Scenarios

Scenario 1: Finding inspiration for daily topic selection

Role: Xiaohongshu creator

Needs: Doesn't know what to write today, wants to see what others in the same niche are posting

How to use：

  1. Enter your content category, such as "Makeup and Beauty"
  2. Browse the TOP50 list to see which topics received high engagement
  3. Refer to the title templates and content features in the viral pattern analysis, and recreate them in your own style

Scenario 2: Competitor Content Monitoring

Role: Content Operations / Brand

Need: Understand what content low-follower dark horses in a certain niche are creating, and judge trends

How to Use：

  1. Select a target category and query it regularly
  2. Pay attention to changes in the content topic distribution and title characteristics of notes
  3. Compare pattern analyses from different dates to discover emerging topic directions

Scenario 3: Viral Pattern Research

Role: Content Strategy Analyst

Need: Systematically study the viral patterns of a certain category

How to Use：

  1. Query the same category for multiple consecutive days
  2. Compare changes in the proportion of title types and content topics across different dates
  3. Export the file package for archiving, making it easier to organize analysis reports

6. Project Architecture

6.1 Directory Structure

xiaohongshu-lowtop/
├── SKILL.md
├── scripts/
│   ├── fetch_explosive_articles.py      # 主脚本：数据获取（原生 socket+SSL） + 分析 + md输出
│   └── generate_html_on_demand.py       # HTML文件包生成
├── references/
│   └── api-spec.md                      # API接口规范
└── assets/
    └── preview-template.html            # HTML页面模版

6.2 Technical Composition

Component Description
Python 3 Core scripting language
socket / ssl Native network requests
json Data parsing
argparse Command-line arguments
HTML templates File package page rendering

6.3 Data Flow

Main query process: User enters a category keyword → matches the corresponding category → calls the data interface (1 API request) → retrieves note data → sorts by engagement → generates a ranking table + cached JSON → automatically analyzes title characteristics and content topics → outputs a complete report

HTML generation process (energy-saving + automatic display): User selects export → reads cached JSON file (0 API requests) → renders HTML page → Agent immediately uses preview_url to display automatically → informs of the save path

7. FAQ

Query-Related Questions

Q1: What should I do if many of the notes I find are unrelated to the category I had in mind?

A: Category matching is based on keywords and content tags, so cross-category content occasionally gets mixed in. If the results deviate significantly, try a more precise keyword; for example, searching "home renovation" is more accurate than searching "renovation".

Q2: Why isn't the data from today?

A: The data is updated once daily at 19:30. Queries before 19:30 will use the previous day's data, which is the normal update rhythm.

Q3: Why is there no response after entering "View More"?

A: Confirm that the total amount of data exceeds 20 items. If the total is no more than 20 items, the first output is all the content.

Subscription-Related Questions

Q4: After subscribing, when is the daily push?

A: The latest data for the day is automatically pushed daily at 19:30.

Q5: Can I subscribe to multiple categories?

A: Currently, you can subscribe to one category at a time; if you want to switch categories, you need to subscribe again.

File Export Questions

Q6: What does the exported file package look like?

A: It generates a complete Xiaohongshu-style page, with each note in an independent card, including the title, author, engagement data, and content analysis. You can open it directly in a browser or export it as a PDF.

8. Resource Index