小红书评论采集 | Xiaohongshu XHS RedNote Comments Scraper

Pull every comment from any Xiaohongshu (小红书 / XHS / RedNote) note. Paste note links, get one clean row per comment — text, author, likes, reply count, IP location, and timestamp. No cookies, no login, no API key.

中文说明

小红书评论采集工具。粘贴笔记链接(网页链接或 App 分享短链均可),导出该笔记下的全部评论:

• 评论正文、点赞数、回复数
• 评论者昵称 / 用户 ID / 主页链接 / 头像
• IP 属地、发布时间
• 可选:把评论回复一并展开成独立行(带 parentCommentId)

无需 cookie、无需登录、无需自备 API key。 支持导出 JSON / CSV / Excel / HTML,也可通过 Apify REST API 接入 n8n、Make、Zapier 或自建流水线。

适用场景:口碑与舆情分析、产品评价挖掘、竞品差评研究、达人粉丝互动分析、AI 训练语料采集。

Support / 联系我们

Bug、字段缺失、需要更高并发?请在 Actor 页面的 Issues 标签提交,我们通常在 24–72 小时内响应。需要定制字段或 SLA,可通过 Apify 站内聊天联系。

What does Xiaohongshu Comments Scraper do?

Xiaohongshu's comment threads are where the honest signal lives — what people actually think about a product, a brand, or a creator. This Actor turns those threads into a spreadsheet.

Give it one or more note URLs. For each note it pages through the full comment list and emits one dataset row per comment, so the output drops straight into a pivot table, a sentiment model, or a vector store. Optionally it flattens replies into their own rows too.

It reads the public comment surface through a maintained backend — you never supply a Xiaohongshu cookie, and nothing breaks when your session expires.

Why use it?

• One row per comment. No nested JSON to unpack, no post-processing before analysis.
• IP location included. Xiaohongshu exposes commenter IP province/region — useful for regional sentiment and for filtering out obvious spam farms.
• Replies, flattened. Turn on includeReplies and sub-comments arrive as their own rows with parentCommentId, so you can rebuild threads or treat them as flat text.
• Cost is capped by you. maxCommentsPerNote is a hard limit, and you are billed per comment returned — so a run can never surprise you.
• No cookies, no breakage. Actors that need your logged-in cookie stop working every few days. This one doesn't use cookies at all.
• Lightweight and fast. No browser is launched, so runs start in seconds and use little memory.

How to use Xiaohongshu Comments Scraper

1. Open the Actor and click Try for free.
2. Paste one or more note URLs into noteUrls. Both formats work:
   • https://www.xiaohongshu.com/explore/<noteId>
   • an xhslink.com/... share link copied from the app (resolved automatically)
3. Set Max comments per note (default 100) — this is your cost ceiling.
4. Optionally switch on Include replies.
5. Click Save & Start, then download as JSON, CSV, Excel, or HTML.

Want this daily? Add the Actor under the Schedules tab — no code required.

Input

Field	Type	Description
noteUrls	array	Note URLs or xhslink.com share links. Required.
maxCommentsPerNote	integer	Hard cap per note (1–5000). Default 100. Also your cost cap.
includeReplies	boolean	Flatten replies into their own rows. Default false.

{
    "noteUrls": [
        { "url": "https://www.xiaohongshu.com/explore/69fba039000000003803426d" }
    ],
    "maxCommentsPerNote": 200,
    "includeReplies": true
}

Output

One row per comment:

{
    "commentId": "6a1b2c3d000000000f00a1b2",
    "noteId": "69fba039000000003803426d",
    "noteUrl": "https://www.xiaohongshu.com/explore/69fba039000000003803426d",
    "content": "太实用了，已经收藏！",
    "likes": 42,
    "replyCount": 3,
    "authorName": "用户A",
    "authorId": "5e2a3c1d000000000100abc7",
    "authorAvatar": "https://sns-avatar-qc.xhscdn.com/avatar/....jpg",
    "authorUrl": "https://www.xiaohongshu.com/user/profile/5e2a3c1d000000000100abc7",
    "ipLocation": "上海",
    "publishedAt": "2026-04-13T09:42:00.000Z",
    "isReply": false,
    "parentCommentId": null,
    "scrapedAt": "2026-08-10T01:12:00.000Z"
}

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

Field	Description
commentId	Xiaohongshu's internal comment ID. Use it to deduplicate across runs.
noteId / noteUrl	The note this comment belongs to.
content	Comment text.
likes	Like count on the comment.
replyCount	Number of replies under this comment.
authorName	Commenter's display name.
authorId / authorUrl	Commenter's user ID and profile link.
authorAvatar	Avatar image URL.
ipLocation	IP province / region as shown by Xiaohongshu.
publishedAt	ISO 8601 comment timestamp.
isReply	true for rows produced by includeReplies.
parentCommentId	Set on reply rows; null on top-level comments.
scrapedAt	ISO 8601 collection timestamp.

Fields Xiaohongshu does not return for a given comment come back as null rather than being silently dropped. In practice commentId, content, likes, replyCount, authorName, authorId, authorAvatar and publishedAt are populated for essentially every comment, while ipLocation is only present on a subset (Xiaohongshu does not attach it to every comment) and image-only comments have an empty content.

How much does it cost to scrape Xiaohongshu comments?

You pay per comment returned — see the Pricing tab on this Actor for the current rate. Because maxCommentsPerNote is a hard cap, your maximum spend per run is simply notes x maxCommentsPerNote x rate.

Platform usage is included — no separate compute or proxy bill. This Actor runs without a browser, so it is cheap and fast to execute.

Free vs. paid Apify plans

Apify free-plan users get 2 sample runs capped at 10 comments each so you can check the data before committing. Further free-plan runs are blocked with an upgrade prompt — this Actor depends on a paid backend, so unlimited free use isn't something we can subsidise. Any Apify paid plan lifts the cap.

Tips

• Deduplicate by commentId when you re-run the same note over time; the Actor already dedupes within a single run.
• Start with maxCommentsPerNote: 50 to sanity-check a note's comment quality before pulling thousands.
• Leave includeReplies off unless you need them — replies count toward your limit and most sentiment work only needs top-level comments.
• Group notes per run. One run with 20 note URLs is cheaper in overhead than 20 separate runs.
• Filter by ipLocation to build region-by-region sentiment views.

FAQ

Do I need a Xiaohongshu account or cookie? No. Nothing to log into, nothing to expire.

Can it get replies to comments? Yes — set includeReplies: true. Replies arrive as their own rows with parentCommentId pointing at the parent, so you can rebuild the thread or analyse everything flat.

What if a note is private or deleted? That note is skipped with a warning and the run continues with the rest — one bad URL never kills the batch.

Why did I get fewer comments than maxCommentsPerNote? The note simply has fewer comments, or Xiaohongshu stopped serving more. The cap is a ceiling, not a target — you are only billed for rows actually returned.

Is scraping Xiaohongshu legal? This Actor collects publicly visible, non-private content — what any visitor can see. How you use the output is your responsibility; comply with your local law and Xiaohongshu's Terms of Service, and don't use the data to identify or profile individuals.

Can I get the data via API? Yes. Every run exposes a stable Apify REST API for input, dataset, and run status.

I need other Xiaohongshu data too. Our 小红书 API | Xiaohongshu Pro Scraper covers keyword search, note details, creator profiles, topic feeds, hot lists, and keyword suggestions.