Xiaohongshu / RedNote Scraper & Intelligence

Not another generic RedNote scraper. This Actor turns Xiaohongshu notes into evidence-linked content opportunities, creator audits, competitor comparisons, consumer-voice reports, and market maps.

Intelligence modes

• demo: run a stable no-Cookie example before supplying any data
• topicRadar: rank proven content examples and suggest actionable angles
• consumerVoice: identify consumer questions, concerns, comparison criteria, mentioned entities, and evidence-backed content directions
• marketMap: map named brands, products, ingredients, devices, medicines, foods, and lifestyle alternatives inside the supplied discussions
• creatorAudit: calculate a creator's engagement baseline, breakout rate, content themes, strongest notes, and data-quality warnings
• brandCompare: compare two to five brands by content share, engagement share, breakout rate, themes, and differentiators

All intelligence modes accept existing note JSON exported from other Actors. Consumer voice and market map also accept existing comment JSON or plain comment strings. Customers can keep their preferred data provider and buy only the analysis. They can alternatively provide their own Xiaohongshu Cookie and collect notes through this Actor.

MVP boundary

Collection modes:

• search: collect notes matching a keyword
• userPosts: collect a creator's public posts
• note: collect one note by URL or ID
• snapshot: parse supplied HTML snapshots or attempt public note pages

The first three modes require the customer's own Xiaohongshu Cookie, entered in Apify's secret input field. The Cookie is used only for the run and is never written to logs or datasets. Residential proxy configuration is supported and recommended for authenticated runs.

The Actor is read-only. It does not like, comment, follow, publish, bypass CAPTCHAs, or attempt to defeat access controls.

Topic radar without a Cookie

{
  "mode": "topicRadar",
  "query": "防晒",
  "uploadedNotes": [
    {
      "noteId": "note-1",
      "title": "敏感肌防晒避坑",
      "url": "https://www.xiaohongshu.com/explore/note-1",
      "creatorName": "示例作者",
      "likes": 5200,
      "comments": 320,
      "collects": 2100,
      "shares": 180,
      "hashtags": ["防晒", "敏感肌"]
    }
  ],
  "maxOpportunities": 10
}

Uploaded data is processed for the current run and is not treated as a login credential.

Consumer voice and market map

Use consumerVoice when the question is "What are customers really discussing, asking, and worrying about?" Use marketMap when the question is "Which brands, products, ingredients, and alternatives exist in the customer's mind?"

{
  "mode": "consumerVoice",
  "query": "血糖管理",
  "regulatedTopic": true,
  "knownEntities": [
    "客户品牌",
    "竞品甲",
    "竞品乙",
    "某种成分",
    "血糖仪",
    "饮食管理"
  ],
  "uploadedNotes": [
    {
      "noteId": "note-1",
      "title": "餐后血糖管理记录",
      "description": "记录饮食、运动和产品选择时遇到的问题。",
      "likes": 800,
      "comments": 120,
      "collects": 350
    }
  ],
  "uploadedComments": [
    {
      "commentId": "comment-1",
      "noteId": "note-1",
      "text": "这个成分和竞品甲有什么区别？价格贵不贵？"
    },
    {
      "commentId": "comment-2",
      "noteId": "note-1",
      "text": "我更想先调整饮食和运动，什么时候需要咨询医生？"
    }
  ]
}

For regulated health and other sensitive topics, claim-risk output is a screening aid only. It does not provide medical, legal, advertising, or regulatory approval.

What it returns

Collection modes return one normalized row per successful note. Intelligence modes return report rows with source-linked evidence.

Depending on the selected mode, reports include:

• evidence-linked consumer questions and discussion dimensions
• named brand, product, ingredient, and alternative mention counts
• comparison dimensions such as results, safety, ingredients, price, usage, experience, trust, and alternatives
• content and engagement share across competitor datasets
• creator baselines, breakout rates, dominant themes, and strongest notes
• content directions tied to observed evidence
• coverage warnings when too few notes or no comments were supplied
• claim-risk matches and safer framing for regulated topics

Outputs describe patterns in the supplied dataset. They do not infer private attributes, diagnose individuals, or guarantee that a sample represents the whole market.

Example input

{
  "mode": "search",
  "cookie": "a1=...; web_session=...; webId=...;",
  "query": "防晒",
  "brandKeywords": ["防晒", "敏感肌"],
  "maxItems": 100,
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"]
  }
}

For local fixture testing:

{
  "mode": "snapshot",
  "htmlSnapshots": [
    {
      "url": "https://www.xiaohongshu.com/explore/demo-note",
      "html": "<html>...</html>"
    }
  ],
  "brandKeywords": ["防晒"]
}

For creator posts:

{
  "mode": "userPosts",
  "cookie": "a1=...; web_session=...; webId=...;",
  "userUrlOrId": "https://www.xiaohongshu.com/user/profile/USER_ID",
  "maxItems": 50
}

Output example

{
  "noteId": "demo-note",
  "url": "https://www.xiaohongshu.com/explore/demo-note",
  "title": "敏感肌防晒实测",
  "creatorName": "示例作者",
  "likes": 1280,
  "comments": 86,
  "collects": 402,
  "shares": 31,
  "matchedKeywords": ["防晒", "敏感肌"],
  "discoveryScore": 87,
  "scoreReasons": [
    "strong public engagement",
    "multiple requested keywords matched",
    "substantive note text"
  ],
  "sourceType": "public-note-page"
}

Local development

$python3 -m unittest discover -s tests -v

To run as an Actor after installing dependencies:

pip install -r requirements.txt
python3 -m src

Monetization recommendation

The Actor uses pay-per-event pricing:

Event	Price
Successfully parsed note	$0.00599
Ranked content opportunity	$0.05
Creator performance audit	$0.49
Two-to-five-brand comparison	$1.49
Consumer voice report	$1.49
Consumer market map	$1.49

Blocked, empty, duplicated, and invalid collection results are not charged. Platform usage and Apify's small automatic events may still appear separately in the run cost summary.

Responsible use

Only collect data you are authorized to process. Follow applicable platform terms and privacy, advertising, and data-protection laws. Do not use the output for harassment, sensitive profiling, or automated decisions about individuals. For health topics, do not present supplements or foods as treatments, imply diagnosis, promise outcomes, or advise users to replace professional care or medication.