Xiaohongshu Scraper: Category Feeds, Notes & Creators

🏆 Part of The Mine Works' 149-actor portfolio · 40,000+ runs · 200,000+ results delivered across our full distribution stack · 99.9%+ success rate

💰 $0.90 / 1,000 results.

⚡ 11 category feeds, not just the Explore page. No login, no browser, no API key. 💸 Pay per note, per full-detail note, or per creator profile. You pick the depth per run.

Scrape Xiaohongshu, also known as RED or RedNote, the Chinese lifestyle platform with roughly 300 million monthly users that drives most beauty, fashion, travel and food discovery in China. No login, no browser, no API key.

✅ 11 category feeds, not just one · ✅ Full note text and topic tags · ✅ Creator profiles with follower counts · ✅ Engagement numbers parsed from 万/亿 into real integers · ✅ No note or creator charge on an empty run · ✅ MCP-ready for AI agents

Scrape the Xiaohongshu discovery feed by category

Pick the Explore feed, or any of Xiaohongshu's own category channels: Fashion, Food, Beauty, Film & TV, Career, Relationships, Home, Gaming, Travel, Fitness and Video. Select several and the run rotates between them, so you get a share from each instead of filling up on the first.

{
  "mode": "feed",
  "categories": ["homefeed.cosmetics_v3", "homefeed.fashion_v3"],
  "maxItems": 100
}

Each note comes back with its title, format (image post or video), author, like count, cover image and a direct note URL.

Get the full text, topics and video URL of every note

Turn on Fetch full note details and each note is opened to add its body text, topic tags, comment / share / save counts, every image URL, the playable video URL and the publish date.

{
  "mode": "feed",
  "categories": ["homefeed.travel_v3"],
  "includeDetails": true,
  "maxItems": 40
}

This costs one extra request per note, so it is billed at the higher per-note rate. Leave it off when you only need a headline scan of what is trending.

Research creators and influencers

Paste Xiaohongshu profile URLs and get each creator's profile plus their recent posts, in one record set.

{
  "mode": "creators",
  "creators": [
    "https://www.xiaohongshu.com/user/profile/5bc41ce0c2ffcc00015326e0?xsec_token=ABxxxx",
    "https://www.xiaohongshu.com/user/profile/555abb13e4b1cf27444532b0?xsec_token=ABxxxx"
  ]
}

Use the full URL you copied from Xiaohongshu, including its xsec_token. Tested across eight creators, profile requests carrying the token succeeded 8 times out of 8, while bare user IDs succeeded only 5 times out of 8 and gave different answers on repeat attempts. A bare ID is retried automatically, but it is not reliable and is not what this mode is built around.

What data do you get

Note records (feed mode)

Field	Description
📝 title	The note's display title
🔗 note_url	Direct link to the note on Xiaohongshu
🎬 note_type	normal for an image post, video for a video
🏷️ category	Which feed the note came from
👤 author_name, author_id, author_url	The creator who posted it
❤️ liked_count	Likes as an integer, plus liked_count_raw as Xiaohongshu wrote it
🖼️ cover_url, cover_width, cover_height	Cover image
⏱️ video_duration_sec	Video length, on video notes

Added when includeDetails is on

Field	Description
📄 description	The note's body text
#️⃣ topics, topic_ids	Topic tags attached to the note
💬 comment_count, share_count, collected_count	Full engagement set
🖼️ image_urls	Every image in the post
▶️ video_url	Playable video file URL
📅 published_at, last_updated_at	ISO timestamps

Creator records (creators mode)

Field	Description
👤 nickname, red_id	Display name and the public RED handle
👥 followers, following, likes_and_collects	Audience size as integers, with the original bucketed strings kept alongside
📍 ip_location	The province Xiaohongshu shows for the account
🎯 profession_tags	The creator's declared niche, e.g. beauty blogger, music blogger
📝 bio	Profile description
📌 recent_notes_returned	How many of their recent posts came with this profile

Real record from a live run

{
  "record_type": "note",
  "note_id": "6a3c95cf000000001603f89b",
  "note_url": "https://www.xiaohongshu.com/explore/6a3c95cf000000001603f89b?xsec_token=AB4GDuBmjOd...",
  "note_type": "video",
  "title": "原来这种情况叫喜金水",
  "category": "Explore (all)",
  "author_name": "汤饭小当家",
  "author_id": "65ef199d0000000005009db2",
  "liked_count": 5934,
  "liked_count_raw": "5934",
  "cover_url": "http://sns-webpic-qc.xhscdn.com/...",
  "detail_fetched": false,
  "scraped_at": "2026-08-30T19:00:00.000Z"
}

About the numbers. Xiaohongshu publishes engagement as localised text: 4.4万 means 44,000 and 1.2亿 means 120,000,000. Every count is returned twice, once parsed to a real integer you can sort and total, and once as the original string so you can audit it.

How the feed actually works, and what that means for you

Xiaohongshu has no feed cursor. Each request returns a fresh sample of its recommendation feed, so this actor keeps requesting until it reaches your maxItems. In a live test, four sequential requests returned 28, 33, 27 and 28 notes with zero overlap, giving 115 unique notes. Duplicates are filtered inside the run and never billed, and once a feed starts returning only notes you already have, that feed stops rather than paying to fetch them again.

The practical consequence: this is a sampler of what Xiaohongshu is actually recommending right now, not a complete archive of a category. That is what makes it useful for trend work.

What this does not do, and why

There is no keyword search. Xiaohongshu's search page returns a valid response, but the results are not in it: they arrive from a signed internal API that requires request signatures computed inside browser JavaScript. A search mode would return zero results on every run, so it is not offered rather than shipped as a feature that never works.

Creator posts have no individual URLs. Xiaohongshu removes note IDs from profile pages entirely, so a creator's recent posts come back with title, format, likes, cover image and publish date, but note_url is null. This was verified directly rather than assumed. For notes with URLs, use feed mode.

What does Xiaohongshu scraping cost here

Event	Price	You pay when
Note scraped	$0.0009	A note is returned from a feed
Note detail scraped	$0.003	A note is returned with its body text, topics and full engagement
Creator profile scraped	$0.006	A creator profile is returned, including their recent posts at no extra charge
Actor Start	$0.005	The run starts. One event per GB of memory, and this actor runs on 512 MB, so one per run

Flat pricing across every Apify plan, not tiered. A run that returns nothing costs the actor start event and nothing else, no note or creator charges. If a detail fetch fails, you still get the note and are billed at the cheaper base rate, never the detail rate. There is no subscription.

Common use cases

China market trend research. Watch what is surfacing in Beauty or Fashion this week before committing to a campaign.

Influencer discovery and vetting. Pull follower counts, niche tags, location and bio for a shortlist of creators, from public profile data.

Competitive and category monitoring. Track which formats and topics are getting engagement in your category.

Cross-platform social listening. Add Chinese social signal alongside Western platforms most tools already cover. Pairs with Douyin Trending & Hashtag Scraper.

Run it on a schedule

1. Run the actor once with the input you want repeated, then click Save as a task.
2. Turn on Monitor mode so later runs return only what is new.
3. In the Apify Console go to Schedules → Create new, name it, set your timezone and a frequency.
4. Under Actors or tasks to run, add the task you saved.

Nothing is charged just for a schedule existing. Full options are in Apify's Schedules documentation.

Requirements

This actor needs Apify's residential proxy. Xiaohongshu does not serve datacenter IP addresses at all, so a run without residential proxy access returns nothing rather than partial data.

FAQ

Do I need a Xiaohongshu account or cookies? No. Everything here comes from public pages rendered without a login.

Why do repeated runs return different notes? Because Xiaohongshu's feed is a live recommendation sample, not a fixed list. That is the platform's behaviour, and it is why monitor mode is worth turning on for scheduled runs.

Can I scrape a specific hashtag or topic? Not directly. Topic tags are returned on every note when includeDetails is on, so you can filter by topic after the fact, but Xiaohongshu's topic search is behind the same signed API as keyword search.

Is it in Chinese? The content is, since it is a Chinese platform. Titles, body text, topics and creator names come back in the original Chinese, ready to pass to any translation step.

Is this GDPR compliant? It collects only what a creator has published publicly on their own profile and posts. You are responsible for having a lawful basis for processing it.

Part of the China social suite

Chinese platforms are mostly unclaimed ground in English-language scraper communities, where the discussion rarely gets past Twitter/X and Instagram. Pair this with Bilibili Scraper for video search and creator stats, and Weibo Scraper for profile stats and post timelines on China's biggest microblog.

Use from Claude, ChatGPT and any MCP agent

https://mcp.apify.com/?tools=themineworks/xiaohongshu-explore-scraper

Or call it programmatically:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });

const run = await client.actor('themineworks/xiaohongshu-explore-scraper').call({
    mode: 'feed',
    categories: ['homefeed.cosmetics_v3'],
    maxItems: 50,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);

Keywords: Xiaohongshu scraper, RED scraper, RedNote scraper, Xiaohongshu API, Xiaohongshu creator profiles, Little Red Book scraper, China lifestyle app scraper, Xiaohongshu MCP server