RedNote Comments Scraper - Xiaohongshu Comments & Replies

RedNote Comments Scraper: Xiaohongshu Comments, Replies, Likes & IP Region

Scrape every top-level comment on a RedNote (Xiaohongshu, 小红书) note from its link or ID: 707 comments and replies from three notes came back in under 9 seconds in our last test run. Each row carries the comment text, the author and their profile link, likes, the true reply count, the posting time, the IP region RedNote shows (in English and in Chinese), attached pictures and @-mentions. No login, no cookies, no xsec_token. Export to CSV, JSON, Excel, or XML.

RedNote's desktop site wants a login, or at least the note's xsec_token, before it loads a comment thread, and a note link copied without that token often opens an error page. This Actor reads the comment feed RedNote serves to its mobile web page, so a bare 24-character note ID is enough, and it returns one flat row per comment or reply that you can load straight into a spreadsheet, a BI tool or a sentiment model.

It reads every top-level comment of every note you give it (482 of 482 in the test above, across notes with 35, 130 and 317 top-level comments), plus the replies RedNote puts under each one. Each row has 27 fields.

Who uses it	What they read RedNote comments for
Brand and social listening teams	What Chinese consumers say under product reviews, launches and KOL posts
China-market researchers	Unfiltered opinions on a category, grouped by the commenter's province
KOL and agency managers	Whether a sponsored post's comments are real engagement or empty praise
Product and CX teams	Complaints and feature requests hidden in comment threads, by keyword
Data scientists	Chinese-language comment datasets for sentiment, topic and intent models

What it does

The Actor takes RedNote note links, share links or note IDs and returns one row per comment and per reply.

• 💬 Every top-level comment: it follows RedNote's comment feed to the end, 20 comments a page, until the note has no more.
• ↩️ Replies as their own rows: the replies RedNote shows under each comment (up to 3, the ones it puts first), linked to their thread with rootCommentId and replyToCommentId. Every comment row also carries its true replyCount.
• 👤 Author details: nickname, user ID, profile link and avatar, plus RedNote's AI-agent flag.
• ❤️ Engagement: likes as a number and as RedNote displays them.
• 📍 IP region: the province or country RedNote shows next to each comment, translated to English, with the original Chinese kept alongside.
• 🖼 Pictures and mentions: original-size links to images attached to a comment, and every @-mentioned user with their profile link.
• 📝 Note context on every row: note ID and link, the note's total comment count and its top-level comment count, and the comment's position in the note.
• 🔎 Filters: keywords, minimum likes, posted-after date and picture-only. Comments that do not match are skipped and not charged.

What you can do with RedNote comment data

🧭 Listen to Chinese consumers in their own words.

Paste the notes that review your product or a competitor's, filter by keywords like 价格 (price) or 质量 (quality), and read the complaints and praise without opening the app.

📍 Map opinion by region.

Group rows by ipLocation to see which provinces talk about a product, and whether overseas Chinese buyers (the United States, Malaysia, Japan) react differently.

📈 Audit a KOL campaign.

Compare noteCommentCount, likeCount and replyCount across a creator's sponsored notes, and read the top comments to see whether the engagement is real.

🤖 Build a Chinese comment dataset.

Collect thousands of comments with their author, time and region for sentiment, topic or intent models. RedNote stickers stay in the text as codes like [笑哭R] (laughing to tears), so emotion is not lost.

Why choose this scraper

	What you get
No login, no token	A bare note ID works. You do not need a RedNote account, cookies or the xsec_token from the share link.
Every top-level comment	The comment feed is read to the end, not only the first screen.
Flat rows, linked threads	One row per comment or reply, so each is its own spreadsheet line, with rootCommentId to rebuild the thread.
IP region in English	Province and country names translated, with RedNote's Chinese label kept for exact matching.
Filters that cut cost	Keywords, likes, date and pictures run as comments are read. Only matching comments are written and billed.
Honest totals	Every row carries the note's own comment counts, so you always know your coverage.

How it compares

All the scrapers below read public RedNote comments from note links. The comparison uses each rival's public input options, README and Store price on September 28, 2026.

Feature	ParseForge	zen-studio/rednote-comments-scraper	khadinakbar/xiaohongshu-comments-scraper	memo23/rednote-comments-scraper
Works from a bare note ID, no login	Yes	Yes	Yes	No: needs the xsec_token, and your login cookie for more than a ~5-comment preview
Row shape	One row per comment or reply	One row per comment, replies nested inside	One row per comment or reply	One row per comment, up to 5 replies nested inside
Replies	Up to 3 per comment (RedNote's public view), plus the true reply count	Yes, on paid plans	The replies expanded on the public page	Up to 5 per comment, plus the reply count
Keyword, likes, date and picture filters	Yes	No	No	No
IP region	English and Chinese	Yes	Yes	Chinese
Price per 1,000 results	$2.49, down to $2.22	$5.99, down to $4.99	$3.00	$5.00
Run start fee	$0.005	$0.05 per GB	$0.00005	$0.005

What a comment looks like

One unedited row from a verified run on September 28, 2026:

{
  "imageUrl": "https://sns-avatar-qc.xhscdn.com/avatar/1040g2jo31uu4q7413q005qca0u389c1cvfo51i0",
  "commentId": "69dc63550000000015004853",
  "type": "Comment",
  "text": "千万不要点进作者的主页，不然就出不来了[捂脸R][捂脸R]",
  "authorName": "晚风轻吟",
  "authorId": "698a0786000000002100b02c",
  "authorUrl": "https://www.xiaohongshu.com/user/profile/698a0786000000002100b02c",
  "authorIsAiAgent": "No",
  "likeCount": 130,
  "likeCountText": "130",
  "replyCount": 12,
  "publishedAt": "2026-04-13T03:30:29.000Z",
  "ipLocation": "Chongqing",
  "ipLocationRaw": "重庆",
  "hasPictures": "No",
  "pictureCount": 0,
  "rootCommentId": "N/A",
  "replyToCommentId": "N/A",
  "position": 1,
  "noteId": "69db721a00000000220290a8",
  "noteUrl": "https://www.xiaohongshu.com/explore/69db721a00000000220290a8",
  "noteCommentCount": 449,
  "noteTopLevelCommentCount": 317,
  "mentionedUsers": [],
  "pictureUrls": [],
  "scrapedAt": "2026-09-28T19:32:25.056Z",
  "error": null
}

More real rows

A reply in that thread. rootCommentId points to the comment above:

{
  "imageUrl": "https://sns-avatar-qc.xhscdn.com/avatar/1040g2jo31q83819h7u005pdi1p9kpa9hcpnh30g",
  "commentId": "69dc6a7d000000001500ab19",
  "type": "Reply",
  "text": "不信你的话，点进去第一反应就是咽了几口口水[暗中观察R]",
  "authorName": "八方来财",
  "authorId": "65b20e53000000001300a931",
  "authorUrl": "https://www.xiaohongshu.com/user/profile/65b20e53000000001300a931",
  "authorIsAiAgent": "No",
  "likeCount": 103,
  "likeCountText": "103",
  "replyCount": 0,
  "publishedAt": "2026-04-13T04:01:02.000Z",
  "ipLocation": "Guangxi",
  "ipLocationRaw": "广西",
  "hasPictures": "No",
  "pictureCount": 0,
  "rootCommentId": "69dc63550000000015004853",
  "replyToCommentId": "69dc63550000000015004853",
  "position": 1,
  "noteId": "69db721a00000000220290a8",
  "noteUrl": "https://www.xiaohongshu.com/explore/69db721a00000000220290a8",
  "noteCommentCount": 449,
  "noteTopLevelCommentCount": 317,
  "mentionedUsers": [],
  "pictureUrls": [],
  "scrapedAt": "2026-09-28T19:32:25.056Z",
  "error": null
}

A picture-only comment with 87 replies. Picture comments have an empty text:

{
  "imageUrl": "https://sns-avatar-qc.xhscdn.com/avatar/1040g2jo31lbsliua5m6g5o1boso0bpub23djkvg",
  "commentId": "6a40e8e9000000002b02bd00",
  "type": "Comment",
  "text": "",
  "authorName": "睡睡睡",
  "authorId": "602bc730000000000101e7cb",
  "authorUrl": "https://www.xiaohongshu.com/user/profile/602bc730000000000101e7cb",
  "authorIsAiAgent": "No",
  "likeCount": 352,
  "likeCountText": "352",
  "replyCount": 87,
  "publishedAt": "2026-06-28T09:27:06.000Z",
  "ipLocation": "Shanghai",
  "ipLocationRaw": "上海",
  "hasPictures": "Yes",
  "pictureCount": 1,
  "rootCommentId": "N/A",
  "replyToCommentId": "N/A",
  "position": 1,
  "noteId": "6a40dd42000000000f02a91b",
  "noteUrl": "https://www.xiaohongshu.com/explore/6a40dd42000000000f02a91b",
  "noteCommentCount": 481,
  "noteTopLevelCommentCount": 130,
  "mentionedUsers": [],
  "pictureUrls": [
    "https://sns-img-qc.xhscdn.com/comment/1040g4b8321v2ll4k7udg5o1boso0bpub5nch2k8"
  ],
  "scrapedAt": "2026-09-28T19:31:54.173Z",
  "error": null
}

Replies that tag someone carry them in mentionedUsers, for example [{"userId": "6a40f5dc000000000e03d001", "nickname": "重生之做五百个美妆项目", "profileUrl": "https://www.xiaohongshu.com/user/profile/6a40f5dc000000000e03d001"}].

Configure the run

Paste note links, share links or IDs into Note URLs or IDs. Links with or without xsec_token, the older /discovery/item/ links, xhslink.com share links and whole copied share messages all work. The Input tab lists every option.

Every comment on two notes, with their replies:

{ "noteUrls": ["https://www.xiaohongshu.com/explore/6a40dd42000000000f02a91b", "69db721a00000000220290a8"], "maxItems": 5000 }

The first 50 top-level comments of a note, without replies:

{ "noteUrls": ["69db721a00000000220290a8"], "maxCommentsPerNote": 50, "includeReplies": false }

Only well-liked comments from the last month that mention price:

{ "noteUrls": ["69db721a00000000220290a8"], "keywords": ["价格", "贵"], "minLikes": 5, "postedAfter": "30 days" }

On a test note with 449 comments (411 rows unfiltered), minLikes: 10 kept 6 rows, onlyWithPictures kept 12, keywords: ["主页", "哈哈"] kept 13, and postedAfter: "2026-05-01" kept 44.

Pricing

Pay-per-event: $0.00249 per comment or reply, dropping to $0.00222 at higher Apify plans, plus $0.005 per run start. You pay only for rows written to your dataset; error rows and comments skipped by your filters are free.

Rows collected	Approximate cost
20	$0.05
1,000	$2.50
10,000	$24.91

New Apify accounts start with $5 in free credit.

Free users

Free-plan runs return up to 10 rows as a preview. Upgrade your Apify plan to collect up to 1,000,000 rows per run.

Run it

1. Create a free Apify account with $5 in credit.
2. Open the RedNote Comments Scraper.
3. Paste one or more note links or IDs into Note URLs or IDs, set Max Items, add any filters, and click Start.
4. Export the results as CSV, Excel, JSON, or XML from the Dataset tab, or schedule the run to follow a note's comments over time.

Run it programmatically through the Apify API or the ApifyClient for JavaScript and Python.

Use with AI agents (MCP)

Give an AI agent live RedNote comments through the Model Context Protocol. Add the Actor to Claude, Cursor, or any MCP client:

$claude mcp add --transport http apify "https://mcp.apify.com?tools=parseforge/rednote-comments-scraper"

Then prompt it in plain language:

• "Read every comment on this RedNote note and summarize the top complaints in English."
• "Which provinces do the commenters on these three notes come from? Give me a count per region."
• "Find the comments on this note that mention price, and list the ten most liked."

Copy this into ChatGPT, Claude, or Cursor to start:

Use the Apify Actor "parseforge/rednote-comments-scraper" to collect comments from RedNote (Xiaohongshu) notes. Input: { "noteUrls": ["<note link or 24-character note ID>"], "maxItems": <n>, "includeReplies": true, "keywords": ["<optional word>"], "minLikes": <optional n>, "postedAfter": "<optional date or span>" }. Rows come back with type (Comment or Reply), text, authorName, authorId, likeCount, replyCount, publishedAt, ipLocation, pictureUrls, mentionedUsers, rootCommentId and noteUrl. Call it with the ApifyClient and my APIFY_TOKEN.

Troubleshooting

Why did a note return no comments?

The note has no comments, was deleted or made private, or the ID is wrong. The run writes a free NO_COMMENTS row for that note and moves on to the next one.

Why are some replies missing?

RedNote's public comment view shows at most 3 replies under each comment, the ones it ranks first. The full reply list needs a signed-in account, which this Actor does not use. replyCount on each comment is the true total, so you can see how much of a thread you have.

Why is text empty?

The comment is a picture only. Its image is in pictureUrls.

Can I sort comments by newest or most liked?

Comments come in RedNote's own order, the one its note page shows. Sort the dataset by publishedAt or likeCount after the run, or use postedAfter and minLikes to keep only what you need.

Why fewer rows than I asked for?

maxItems counts comments and replies together across all notes, and your filters skip comments that do not match. A note with 300 top-level comments and few replies returns about 400 rows.

FAQ

Question	Answer
Do I need a RedNote account, cookies or an API key?	No. It reads the comments RedNote serves to logged-out visitors of its mobile web page.
Do I need the xsec_token from the share link?	No. A bare 24-character note ID works, and links with or without the token are accepted.
Does it get every comment?	Every top-level comment, and up to 3 replies under each (the ones RedNote shows publicly). Each comment row carries the true reply count.
Can I paste an xhslink.com share link?	Yes, or the whole copied share message. The Actor follows the link to the note.
What is the IP region?	The province or country RedNote displays under each comment, based on the commenter's IP. It is RedNote's label, not a precise location.
Are like counts exact?	They are the numbers RedNote displays. Above 10,000 RedNote abbreviates them (1.2万), and likeCountText keeps that text.
Can I get the note itself (title, text, likes)?	Use the RedNote Notes Scraper for note details, and join on noteId.
How many notes per run?	As many as you like. Notes are read three at a time, and maxItems caps the total rows.
Is this an official RedNote product?	No. It is unofficial and reads only public RedNote data.

Related actors

• RedNote (Xiaohongshu) Notes Scraper: note titles, text, hashtags, likes, saves and galleries.
• Weibo Scraper: Weibo posts, users, hot search and comments.
• Bilibili Comments Scraper: full comment threads and replies from Bilibili videos.
• TikTok Comments Scraper: comments from public TikTok videos.
• YouTube Comments Scraper: YouTube comments with replies and likes.

Browse the full ParseForge collection for more scrapers.

🆘 Need help? Open our contact form with your run ID, your input, and what you expected.

⚠️ Disclaimer. This Actor is unofficial and is not affiliated with, endorsed by, or sponsored by Xiaohongshu (RedNote). It collects only publicly available RedNote data. You are responsible for using the data in compliance with RedNote's terms and applicable laws, including PIPL, GDPR and CCPA. Do not use it to identify, profile, or target individuals.