Xiaohongshu Skills for AI agents

Search Xiaohongshu notes, creators and product reviews, and follow topic trends with live data your agent reads through SandBase.

Built for Xiaohongshu · 17

Data & research · 4

Writing · 4

Engagement · 1

Planning · 5

Review · 3

Output excerpt

Fictional sample data. … 办公室桌面绿植清单 (图文, 11 days) | Stage | Value | Baseline median | Peers | Gap | | reach | 4100 | 10000 | 6 | −59% | | open | 13.0% | 12.5% | 6 | +4% | | engagement | 20.1% | 20.4% | 6 | −2% | | follow | 1.13% | 1.31% | 6 | −14% | Candidate bottleneck: reach, 59% below the median of 6 image notes. Once shown, the note was opened and engaged with like its peers; follows were 14% lower, under the 15% threshold. The peers are 8 to 29 days old. … Hypotheses (not causes; the export cannot show why a note was shown less): … Candidate bottleneck: open, 44% below baseline, but the baseline is weak: only 2 mature video notes to compare with.

Xiaohongshu Note Analytics

Output excerpt

Fictional sample data. … | Measure | Previous | Current | Change | … | Compared notes (7+ days old) | 6 | 4 | | … | Median age of compared notes (days) | 23.5 | 12 | | … | Follows (total) | 86 | 42 | −51.2% | … | Median follows per view | 1.27% | 1.20% | −5.4% | … The totals do not show a drop in how notes performed. … Confounded (the same notes): 图文 = evening; 视频 = midday = Thursday. … | 2026-11-10 | Tue | 20:30 | 图文 | 暖气房里绿植放哪儿 | keep | | 2026-11-12 | Thu | 20:00 | 视频 | 水培绿萝怎么换水 | test | | 2026-11-14 | Sat | 20:30 | 图文 | 琴叶榕掉叶先查三件事 | keep | Test: video at evening instead of midday. … When: rerun with --as-of 2026-11-20 or later, when the test note is 7 days old (checked with the script).

Xiaohongshu Post Review

Output excerpt

Fictional example data. … [title] 全网最好喝的桂花乌龙!喝了7天痘痘少一半 … | 1 | 喝了7天痘痘少一半 (title), 痘痘真的少了 (cover), 脸上的痘痘少了一半 (body) | ADL-17, ADL-28 (law); ED-1 (editorial) | None. E2 records no skin effect; E2 says 5 days, not 7 | Must fix | Delete every acne claim. Do not replace it with a milder skin claim | … | 8 | #不是广告 纯自用分享 | ADL-4, ADL-14, IAM-9, GID-3, GID-6 (law) | E1: gifted, commission, product card | Must fix | Replace with a prominent 【广告】 at the start of title and body, and state gift, card and commission | … [title] 【广告】办公室喝了5天的桂花乌龙 … | Shared runs, n = 8 | 1 (longest 19) | 0 (longest 3) | … Rules checked 2026-11-05. This is a pre-publication review, not legal advice or platform approval.

Xiaohongshu Publish Precheck

Also works with Xiaohongshu · 9

Local Market Research

Research local markets by combining Google Maps reviews, Xiaohongshu local content, Weibo discussions, and web search. Covers customer sentiment by location, local competition analysis, neighborhood trends, and location-specific brand perception.

Product Intelligence

Research products across Amazon, Google Shopping, and Xiaohongshu to understand pricing, reviews, market positioning, and consumer sentiment. Covers competitive product analysis, pricing benchmarks, feature comparison, and customer satisfaction signals.

Review Aggregator

Aggregate and analyze product and business reviews from Google Maps, Amazon, Akta, and Xiaohongshu. Cross-platform review intelligence including rating patterns, common themes, competitive comparison, and customer satisfaction trends.

Audience Research

Research target audiences across Reddit, Twitter, Instagram, and Xiaohongshu to understand demographics, interests, pain points, language patterns, and content preferences. Produces audience personas backed by real community data.

Brand Monitoring

Monitor brand mentions, sentiment, and reputation across Twitter, Reddit, news, and social platforms through SandBase. Use when asked for brand monitoring, reputation tracking, social listening, PR coverage analysis, or competitive brand intelligence.

Hashtag Tracker

Track hashtag performance across Instagram, TikTok, Twitter, and Xiaohongshu simultaneously. Monitors post volume, engagement rates, top contributors, and content patterns for campaign tracking, trend research, and community activation measurement.

KOL Discovery

Discover and evaluate KOLs (Key Opinion Leaders) and creators across TikTok, Instagram, YouTube, and Xiaohongshu through SandBase. Use when asked for influencer research, creator outreach lists, KOL evaluation, or partnership prospecting.

Related Lab tools

THE FIELD GUIDE

A working guide to Xiaohongshu Skills.

Search Xiaohongshu notes, creators and product reviews, and follow topic trends with live data your agent reads through SandBase.

Install the Skill in your agent, connect SandBase and name the keywords, creators or products to research on Xiaohongshu.

Connect your task to a workflow.

Start with a task, then explore the possibilities.

Social media

Create visual stories, short videos and imagery around an idea your audience can connect with.

Explore Social media

Market & competitors

Research markets, competitors, pricing and reviews with live data your agent gathers through SandBase.

Explore Market & competitors
FIND YOUR FIT

Choose your Xiaohongshu workflow.

Compare the inputs and deliverables before you begin.

What you haveWhat to look forNext step
A clear briefInstall the Skill in your agent, connect SandBase and name the keywords, creators or products to research on Xiaohongshu.Browse this collection
A connected agentInstall the Skill with its NPX command and connect your agent to SandBase. API calls are billed to your SandBase organization.Read the Skills guide
A different outputCreate visual stories, short videos and imagery around an idea your audience can connect with.Social media
PUT IT TO WORK

From install to a sourced answer.

  1. Install the Skill

    Copy the NPX command from the Skill page and run it in SandBase Agent, Codex, Claude or Cursor.

  2. Connect SandBase

    Connect your agent to SandBase so the Skill can call its data APIs.

  3. Ask and check

    Describe what you need, then check the sources in the output before you use it.

FAQ

Xiaohongshu Skills: common questions.

A few things to know before you start.

What is an API Skill?

A set of instructions your agent follows to call SandBase data APIs, such as search, social or business data, and turn the results into a sourced answer.

What account do I need, and who pays for the calls?

You need a SandBase account connected to your agent. SandBase API calls are billed to your SandBase organization.

Which agents are supported?

SandBase Agent, Codex, Claude and Cursor. Each Skill page lists its supported agents and the install command.