小红书 API | Xiaohongshu (XHS / RedNote) Pro Scraper

Full-featured 小红书 / Xiaohongshu / XHS / RedNote data API. Seven modes, one Actor, clean structured JSON. No login, no cookies, no broken-token surprises.

Built for serious workloads — market research teams pulling daily competitive intelligence, AI training pipelines harvesting Chinese-language UGC, KOL discovery tools, brand-watch dashboards.

中文说明

小红书数据 API,一个 Actor 覆盖 七种模式:关键词搜索、笔记详情、博主主页与作品、话题(标签)信息流、搜索热榜、关键词联想、公开热门流。

• 无需 cookie、无需登录,不会因为 token 过期而中断。
• 返回扁平化 JSON:标题、正文、作者、点赞、收藏、评论数、话题标签、图集、视频地址、发布时间、IP 归属地。
• 支持导出 JSON / CSV / Excel / HTML / XML,也可通过 Apify REST API 接入 n8n、Make、Zapier 或自建流水线。
• 适用场景:选品与市场调研、品牌舆情监控、KOL 达人挖掘、SEO 关键词研究、AI 训练语料采集。

计费:列表类结果 $5.00 / 1,000 条,笔记详情 $20.00 / 1,000 条,平台计算费用已包含(不额外收取)。详见下方 Pricing。

Support / 联系我们

Bug、字段缺失、需要新的模式或更高并发?请在 Actor 页面的 Issues 标签提交,我们通常在 24–72 小时内响应。需要定制端点、专用代理或 SLA,可通过 Apify 站内聊天联系。

What does this Actor do?

Pick a mode, give it inputs, get a dataset:

• Keyword search (search) — search Xiaohongshu by keyword with full sort and filter options (newest, most liked, video-only, last 24h, etc.)
• Note details (notes) — paste a note URL or an xhslink.com share link, get the full body, all images, hashtags, engagement counts, location, and optional comments.
• User profile + posts (user) — fetch a creator's profile (followers, IP location, bio) plus their recent posts in one run.
• Topic / hashtag feed (topic) — pull every recent note tagged with a topic.
• Search hot list (hotsearch) — the daily curated 热榜 board with rankings and heat scores.
• Keyword suggestions (suggest) — seed a keyword, get Xiaohongshu's autocomplete suggestions. Built for SEO and content research.
• Public trending feed (trending) — the anonymous "Hot Notes" homepage strip, no Xiaohongshu account needed.

Each note returns a flat, schema-validated JSON row that's downloadable as JSON, CSV, Excel, HTML, or XML — and queryable via the Apify REST API for pipeline integration.

Why use Xiaohongshu Pro Scraper?

• No cookies, no breakage. Many Xiaohongshu actors rely on the customer's logged-in cookie, which expires every few days — customers end up with a stream of 401s. We use a maintained mobile-app surface that stays stable.
• Seven modes in one Actor. Search, note detail, user, topic, hot list, keyword suggestions, trending — pick a mode per run; no juggling separate Actors.
• Topic feeds and keyword research included. Hashtag-level feeds and Xiaohongshu autocomplete are rarely available elsewhere in the Store.
• Full-field output. Title, body text, author, likes, collects, comment count, hashtags, image gallery, video URL, publish time, geo / IP location — all in one row.
• AI training-ready. Chinese-language UGC, clean JSON, deduplicable by noteId. Drop straight into a vector DB, an LLM fine-tune set, or a trend classifier.
• Honest per-type pricing. Cheap list results and expensive full-detail results are billed separately, so a 1,000-note keyword sweep costs $5 — not $20.

How to use

1. Open the Actor and click Try for free.
2. In the Input tab, pick a Scrape mode (search is the default).
3. Fill the inputs for your mode:
   • search: keywords, optional sort/filter
   • notes: full note URLs or xhslink.com share links
   • user: user IDs
   • topic: topic names (without #)
   • suggest: seed keywords
   • hotsearch / trending: nothing required
4. Set Max items per input (default 30).
5. Click Save & Start. Results stream into the dataset live.
6. Download or stream the dataset; schedule the run under the Schedules tab.

Input

Field	Type	Used in mode	Description
mode	string	all	One of search, notes, user, topic, hotsearch, suggest, trending.
keywords	array	search	Keywords to search.
noteUrls	array	notes	Note URLs or xhslink.com share links to fetch details for.
userIds	array	user	User IDs to fetch profiles and posts for.
topicNames	array	topic	Topic / hashtag names (without #).
suggestKeywords	array	suggest	Seed keywords for autocomplete suggestions.
maxItemsPerInput	integer (1–500)	all	Cap per input. Default 30.
sortType	enum	search	general / time_descending / popularity_descending / comment_descending / collect_descending.
noteType	enum	search	不限 / 视频笔记 / 普通笔记 / 直播笔记.
timeFilter	enum	search	不限 / 一天内 / 一周内 / 半年内.
fetchComments	boolean	notes, search	Also fetch up to 20 top comments per note. Adds run time.
proxyConfiguration	object	trending only	Only used for the trending fallback.

JSON example — keyword search:

{
    "mode": "search",
    "keywords": ["AI编程", "护肤"],
    "maxItemsPerInput": 100,
    "sortType": "popularity_descending",
    "timeFilter": "一周内",
    "fetchComments": false
}

Output

{
    "noteId": "69fba039000000003803426d",
    "title": "Claude Code 零基础终极教程！",
    "bodyText": "这是你能看到的最完整的Claude Code 零基础系统教程。 🔹 从安装到接入第三方模型 🔹 讲透 Hooks、S…",
    "author": "木子不写代码",
    "authorId": "5e2a3c1d000000000100abc7",
    "likes": 117895,
    "collects": 172097,
    "commentsCount": 4521,
    "isVideo": true,
    "imageUrls": [
        "https://sns-webpic-qc.xhscdn.com/.../1.jpg"
    ],
    "videoUrl": "https://sns-video-bd.xhscdn.com/.../master.mp4",
    "hashtags": ["#ClaudeCode", "#AI编程", "#提示词"],
    "publishedAt": "2026-04-13T09:42:00.000Z",
    "location": "上海",
    "comments": [
        { "author": "用户A", "text": "太实用！", "likes": 42 }
    ],
    "noteUrl": "https://www.xiaohongshu.com/explore/69fba039000000003803426d",
    "keyword": "AI编程",
    "scrapedAt": "2026-05-22T09:26:33.733Z"
}

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel.

Output fields

Field	Description
noteId	Xiaohongshu's internal note ID.
title	Note title.
bodyText	Full note body text.
author	Display name of the author.
authorId	Internal user ID — pair with mode: "user" to drill into the author.
likes	Like count.
collects	Collect (favourite) count.
commentsCount	Total comment count.
isVideo	True if Xiaohongshu marks the note as a video.
imageUrls	All images in the note.
videoUrl	Video src for video notes.
hashtags	Hashtags / topics referenced.
publishedAt	ISO 8601 publish time.
location	Author IP location (city / region).
comments	Up to 20 top comments when fetchComments is on.
noteUrl	Public deep link.
keyword	Search keyword that surfaced this note (search mode only).
topic	Topic name (topic mode only).
userId	Author ID (user mode only).
scrapedAt	ISO 8601 timestamp of scrape.

How much does it cost to scrape Xiaohongshu?

This Actor uses pay-per-event pricing, billed by result type:

Event	Price	Modes
Search / list result	$5.00 / 1,000 ($0.005 each)	search, topic, user, hotsearch, suggest, trending
Full note detail	$20.00 / 1,000 ($0.02 each)	notes

List results are cheap because they are cheap for us to produce. Full note details hit a far more expensive upstream endpoint, so they are priced separately instead of averaging that cost into every row — which is why a 1,000-note keyword sweep costs $5, not $20.

Platform usage is included. You are not billed separately for compute or proxies.

Free vs. paid Apify plans

Apify free-plan users get 2 sample runs capped at 10 results each so you can verify the data shape and quality before committing. Further runs from free accounts are blocked with an upgrade prompt — this Actor relies on a paid backend, so we can't subsidise unlimited free use. Any Apify paid plan (see Apify pricing) unlocks unlimited runs at the rates above.

Cost estimates

Goal	Items / run	Mode	Est. cost / run
Quick scan of one keyword	50	search	~$0.25
Keyword research sweep	200 suggestions	suggest	~$1.00
Hashtag round-up	500	topic	~$2.50
Daily brand monitoring	1,000	search	~$5.00 / day
KOL deep dive (full details)	200 notes	notes	~$4.00

How to keep costs down

Use search, topic, or user mode whenever you can — they already return title, body, likes, collects, comment count, hashtags, and images at the list rate. Only reach for notes mode when you specifically need one note's complete detail record.

Tips

• Schedule, don't poll. Xiaohongshu rotates trending content every few hours; scheduling daily / 6-hourly delivers the right cadence vs cost.
• Deduplicate by noteId. Sequential runs overlap; merge and dedupe on the noteId field.
• Use suggest before search. Mine the autocomplete list first, then search only the terms that matter.
• Pair user + notes modes. Use mode: "user" to find prolific authors, then drill into a specific top note with mode: "notes".
• Pair popularity_descending with timeFilter: "一周内" to surface this week's breakout posts.
• Don't enable fetchComments unless you need them. Each comment-bearing note takes an extra request and adds run time.

FAQ

Q: Why is this Actor different from the other Xiaohongshu scrapers on the Store? A: Most existing actors depend on the customer providing a logged-in Xiaohongshu cookie that expires every few days. We don't. We've engineered around the public mobile-app surface so the Actor stays stable across cookie expiries — and we bill list results separately from full details so you don't overpay for bulk keyword work.

Q: Will this break when Xiaohongshu changes their API? A: Possibly. Our team maintains the backend; we patch usually within 24-72 hours of a backend change. Bug reports via the Issues tab.

Q: Is scraping Xiaohongshu legal? A: We pull only publicly-accessible content — what any visitor can see without an account. Your usage is your responsibility; comply with your jurisdiction's law and Xiaohongshu's Terms of Service. Do not scrape private profiles or use the data to identify individuals.

Q: Can I get my data via API? A: Yes. Every Apify run exposes a stable REST API for inputs, datasets, and run status. Plug into n8n, Make, Zapier, or your own pipeline.

Q: Is there a free tier? A: Apify gives new accounts free platform credit, and this Actor allows 2 sample runs of up to 10 results each on free plans so you can evaluate the output before paying.

Q: How fast are runs? A: Search mode: ~5 seconds per page (≈20 notes). User mode: ~3 seconds per page. Note-detail mode: 1-2 seconds per note. Topic mode: ~4 seconds per page.

Q: I need a richer integration or higher-volume contract. A: Reach out via Apify chat. Custom endpoints, dedicated proxies, and SLAs available on request.