Xiaohongshu (RedNote) Scraper - Notes, Profiles & Comments

What does Xiaohongshu (RedNote) Scraper do?

Xiaohongshu (RedNote) Scraper turns Xiaohongshu (小红书, RedNote, XHS) notes into clean rows: the full note text, every image and video link, likes, saves, comments and shares, hashtags, the publish date, the author, the video transcript and RedNote's own English translation. Give it note links, share links, note ids, creator profiles or explore-feed categories. No login, no cookies, no app.

Also known as: RedNote scraper · Xiaohongshu scraper · XHS scraper · 小红书 scraper · Little Red Book scraper.

• 📝 Full notes — title, text, hashtags and topics, @mentions, publish and last-edit time, IP region.
• 🌐 English translation — titleEnglish, descriptionEnglish and hashtagsEnglish from rednote.com's own translation.
• 🎬 Video + transcripts — every stream (h264/h265, size, bitrate, fps), subtitle tracks, and the spoken text in the original language and in English.
• 🖼️ Images that do not expire — every image as an unsigned full-size CDN file next to the signed page link.
• 👤 Creator profiles — name, RED id, bio, exact likes, saves and notes totals, tags, verification, newest ~30 notes.
• 💬 Comment preview — the comments Xiaohongshu shows signed-out visitors, with replies and who they answer.
• 🔁 Monitor mode — schedule it and get only notes it has not delivered before.
• 🔗 Any link format — xiaohongshu.com and rednote.com links, xhslink.com share links, whole pasted share text, bare 24-character ids.

What data can you get from Xiaohongshu?

🆔 Note id + link (noteUrl)	📝 Title + full text	🌐 English title, text, hashtags
❤️ Likes (likes)	⭐ Saves (collects)	💬 Comment count
🔁 Shares	🔢 Exact or rounded flag (countsExact)	📅 Published + last edited
📍 IP region (ipLocation)	#️⃣ Hashtags + topic ids	👥 @mentions
🖼️ Images (signed + unsigned original)	✨ Live Photos	🎬 Video streams + backups
⏱️ Video length, size, codec, fps	🗣️ Transcript (original + English)	🎞️ Subtitle tracks (.srt)
👤 Author id, name, avatar, link	🏷️ Profile: RED id, bio, gender, tags	📊 Profile: exact likes, saves, notes
✅ Verified account flag	🗂️ Newest ~30 notes per profile	💬 Comments: text, likes, replies, region

How to scrape Xiaohongshu with Xiaohongshu (RedNote) Scraper

1. Create a free Apify account — the free plan includes monthly credit.
2. Open Xiaohongshu (RedNote) Scraper and paste note links, profiles, or pick explore-feed categories.
3. Set filters (note type, minimum likes, posted within, keywords) and add-ons (comments, author profiles).
4. Click Start.
5. Download the rows as JSON, CSV or Excel, or read them through the API.

⬇️ Input

Parameter	Type	Description
noteUrls	array	Note links (xiaohongshu.com / rednote.com), xhslink.com share links or share text, or 24-character note ids.
profileUrls	array	Profile links, xhslink.com profile share links, or 24-character user ids.
feedCategories	multi-select	Explore channels: recommend, fashion, food, cosmetics, movie_and_tv, career, love, household_product, gaming, travel, fitness, video, kpop. Empty input reads recommend.
noteType	select	all (default), image or video.
minLikes	integer	Keep notes with at least this many likes.
postedWithinDays	integer	Keep notes published in the last N days.
includeKeywords / excludeKeywords	array	Keep / drop notes whose title, text or hashtags (Chinese or English) contain a word.
includeComments	boolean	Add the comment preview as comment rows. Default off.
maxCommentsPerNote	integer	1-20 comment rows per note. Default 20.
includeAuthorProfiles	boolean	Add one profile row per note author. Default off.
includeTranscripts	boolean	Add transcript text for videos with subtitles. Default on, no extra charge.
monitorMode / monitorStoreName	boolean / string	Deliver only notes and comments not delivered before under this memory name.
maxItems	integer	Most note + profile rows per run. Default 50. 0 = no cap.
compact	boolean	Core fields only, for AI and MCP pipelines.
includeRawData	boolean	Attach the page's own note or profile object as raw.
proxyConfiguration	proxy	Default: no proxy.
maxConcurrency	integer	Pages fetched at the same time, 1-16. Default 6.

Nothing is required: an empty input reads the Recommended feed.

Input JSON

{
  "noteUrls": [
    "https://www.xiaohongshu.com/explore/6abfc91e000000001403b27c?xsec_token=ABrpeLWo2ZkPrnMon23RDGH6IWtIZLJyYhzChfic_I1As%3D&xsec_source=pc_feed",
    "6a329deb000000000f01dbb4"
  ],
  "profileUrls": ["https://www.xiaohongshu.com/user/profile/6444d090000000001002822f"],
  "feedCategories": ["food"],
  "includeComments": true,
  "maxItems": 30
}

Keep the xsec_token in note links

Xiaohongshu opens most notes only with the xsec_token that comes in the link you copy from the site or app. Paste the whole link, the xhslink.com share link, or the whole share text — the actor finds the link inside it. A bare note id opens when rednote.com publishes that note without a token (it did for 11 of 12 ids we took from rednote.com pages, 1 of 12 from xiaohongshu.com feeds); otherwise it is listed as tokenRequired in the run summary and not charged.

Use feed categories to discover notes

Each explore-feed page load brings 35 new notes (rednote.com); the actor opens every one for its full details. With several categories, maxItems is split evenly between them.

Plan on about 40-55 notes per run

Xiaohongshu limits how many pages one IP address may open while signed out (see Limits). The comment preview and author profiles each add one page per note, so they lower the count. Keep maxItems around the default 50 and run more often instead: a schedule with monitor mode on collects a feed or a list of links run after run without paying twice for the same note.

⬆️ Output

One row per note, profile or comment (recordType). A note row from our test run (trimmed):

{
  "recordType": "note",
  "noteId": "6abfc91e000000001403b27c",
  "noteType": "video",
  "title": "妈妈牌炖肘子",
  "titleEnglish": "Mom's homemade stewed pork knuckle",
  "likes": 3035, "collects": 196, "commentsCount": 121, "shares": 199,
  "countsExact": true,
  "publishedAt": "2026-10-02T15:09:18.000Z",
  "ipLocation": "甘肃",
  "authorName": "2号爱吃肉的比基尼小姐姐",
  "hashtags": ["美食掌门人", "沉浸式吃饭", "猪脚饭"],
  "hashtagsEnglish": ["gourmetmaster", "justvibingwithmyfood", "porkknucklerice"],
  "coverUrlOriginal": "https://sns-img-qc.xhscdn.com/1040g2sg325qvevlc4ee05o4l5q308f9k6ks93p8",
  "videoUrl": "https://sns-v28.rednotecdn.com/stream/79/110/258/01eabfc91e9921074f037001a0fd2e3614_258.mp4",
  "videoDuration": 393, "videoWidth": 720, "videoHeight": 1280,
  "videoSubtitleLanguages": ["en-US", "zh-CN"],
  "transcript": "呼肘子。妈妈牌货肘子\n粉皮白菜肉，还有炒藕片\n…",
  "transcriptEnglish": "Call for pork knuckles. Mom's style pork knuckles\n…",
  "sourceHost": "rednote.com"
}

A profile row and a comment row:

{ "recordType": "profile", "userId": "6444d090000000001002822f", "nickname": "九派新闻", "redId": "63034816395",
  "ipLocation": "湖北", "followersText": "1万+", "followers": 10000, "followersIsExact": false,
  "likesTotal": 23496423, "collectsTotal": 1517468, "notesCount": 15882, "verified": true, "recentNotesCount": 32 }
{ "recordType": "comment", "commentId": "6ac03203000000000801f848", "noteId": "6abfc91e000000001403b27c",
  "content": "不善做饭有别的特长[斜眼R]", "likes": 10, "isReply": true, "repliedToUserName": "龙龙龙龙是杰",
  "authorName": "天涯若比蔺🎸", "ipLocation": "辽宁", "createdAt": "2026-10-02T22:36:51.000Z" }

Fields and measured fill

Measured on our test runs of 2026-10-16 (feed runs of 43, 56 and 87 notes, 20 notes with 20 author profiles and 248 comments, a 5-link run, the 10-row demo).

Fields	Fill
noteId, noteUrl, noteType, likes, publishedAt, authorId, authorName, imageUrls, imageUrlsOriginal, coverUrlOriginal, shareDisabled	100% of notes
title / description	95-100% / 97-100% (some notes are text-only or title-only)
collects, commentsCount, shares	95-100%
titleEnglish, hashtagsEnglish	95-100% of notes read from rednote.com; none from xiaohongshu.com
descriptionEnglish	60-80% (notes whose text is only hashtags have none)
hashtags	85-100%
video fields (videoUrl, streams, length, size, codec, md5, thumbnail)	100% of video notes (30-35% of feed notes are videos)
transcript, transcriptEnglish	50-100% of video notes (videos with RedNote subtitles)
ipLocation	40% of linked notes, 0-11% of feed notes, 85% of profiles, 98-100% of comments
profile redId, notesCount, likesTotal, collectsTotal, followersText, verified	100% of profiles
profile tags / location / profession	75% / 55% / 25% (20 author profiles)
comment content, likes, authorName, createdAt	98-100%
reply repliedToUserName	97% of replies

Not delivered: keyword search results, full comment threads past the first screen, a creator's full note list, and note ids on profile cards (Xiaohongshu shows none of these to signed-out visitors); exact follower counts (shown only as a bucket such as 1万+); the "nice" counter and collection counts (always blank or 0 for signed-out visitors); follow and like state (signed-in only); media files (links only — signed links expire after hours to days, imageUrlsOriginal links do not carry an expiry stamp).

Use cases

• 📈 Brand and KOL marketers — track notes in a category or by a creator: likes, saves, comments, shares and the date.
• 🌏 Cross-border e-commerce teams — read Chinese consumer notes in English with the translated title, text, hashtags and transcript.
• 🕵️ Competitor watchers — schedule note links or a feed with monitor mode on and get only the new notes.
• 🎬 Content researchers — video length, transcripts and subtitles for format analysis.
• 🤖 AI and RAG pipelines — compact rows with bilingual text, over the API or MCP.
• 🧑‍💼 Influencer scouts — exact likes, saves and notes totals per creator, plus RED id, tags and region.

How much does it cost to scrape Xiaohongshu?

Pay per result, no start fee: $0.0009 per note, $0.0009 per profile and $0.0009 per comment row. Transcripts and translations cost nothing extra. Filtered notes, duplicates, notes seen in earlier monitor runs and links that do not open are never charged.

Run	Rows	Cost
10 notes	10	$0.009
50 notes (default cap)	50	$0.045
20 notes + author profiles + comment preview	288 (20 + 20 + 248)	$0.26

Integrations, API, scheduling & MCP

• Schedule a task hourly or daily with monitor mode on to collect only new notes from a feed or a list of links.
• Webhooks, Make, Zapier, n8n, Google Sheets, Slack and Airtable from the Integrations tab.
• API: call it from Python or JavaScript with apify-client, then read the dataset.
• MCP: use it from Claude, ChatGPT or Cursor through the MCP tab; turn on Compact output for shorter rows.

More social media scrapers

🎵 Douyin Scraper Videos, creators, comments, hot search.	💬 TikTok Comments Scraper Comments, replies, likes.	👤 TikTok Profile Scraper Followers, bio, stats.
🗣️ Instagram Reel Transcript Speech to text, no login.	📺 YouTube Transcript Scraper Subtitles with timestamps.	🧵 Threads Profile Scraper Followers, bio links, posts.

See every actor on our profile.

Limits & empty runs

• Pages per IP. Xiaohongshu shows signed-out visitors a human check (a CAPTCHA page) after a number of note pages from one IP address. On fresh datacenter IPs we measured it at the 45th-64th rednote.com page (64 at one page every 0.8 s, 2 s and 5 s alike — speed does not change it), and later the same evening some IPs got it within the first 1-7 pages. The run stops at the first one: rows already saved are kept, nothing after it is fetched or charged, and the CAPTCHA is never answered. Our feed runs delivered 43-56 notes before it. The same IPs were served again 2-28 minutes after their check, so the limit works like a rolling window.
• Comments are the first screen only: up to 5 top comments with up to 3 replies each (up to 20 rows per note). hasMoreReplies marks threads that have more.
• Followers are a bucket for signed-out visitors (1万+ → followers: 10000, followersIsExact: false); likes, saves and note totals are exact.
• Image, video and subtitle links from the page are signed and expire; download them promptly or use the unsigned imageUrlsOriginal.
• Links that need a token, deleted notes and links that are not Xiaohongshu are listed in RUN_SUMMARY and never charged. A run where nothing could be read fails with a clear status message.

❓ FAQ

How much does Xiaohongshu (RedNote) Scraper cost?

$0.0009 per note, profile or comment row, with no start fee. 50 notes cost $0.045.

How many notes can Xiaohongshu (RedNote) Scraper return per run?

As many as Xiaohongshu serves one IP before its human check: 43-56 notes in our feed runs, fewer when the IP was checked recently. Schedule several smaller runs with monitor mode for more.

Does Xiaohongshu (RedNote) Scraper translate notes into English?

Yes. Notes are read from rednote.com, which carries RedNote's own English title, text and hashtags; video transcripts come in the original language and in English. A note that only opens on xiaohongshu.com (the fallback) has no English fields.

Can Xiaohongshu (RedNote) Scraper search notes by keyword?

No. Xiaohongshu shows search results only to signed-in users. Use explore-feed categories with includeKeywords to keep notes that mention your words.

Can I get comments with Xiaohongshu (RedNote) Scraper?

Yes, the preview Xiaohongshu shows signed-out visitors: up to 5 top comments and 3 replies each per note, with likes, region, time and who a reply answers.

Can I schedule Xiaohongshu (RedNote) Scraper to monitor creators or a feed?

Yes. Save a task, add a schedule and turn on Only new notes (monitor mode). Notes delivered before are skipped and not charged.

Can I use Xiaohongshu (RedNote) Scraper with the Apify API or an MCP server?

Yes — call it from Python or JavaScript with apify-client, from Make/Zapier/n8n, or as a tool in Claude, ChatGPT or Cursor over MCP.

Do I need a Xiaohongshu account or cookies?

No. The actor reads the pages any signed-out visitor sees. No login, no cookies, no app.

Why did my Xiaohongshu (RedNote) Scraper run return fewer notes than I asked for?

Xiaohongshu showed its human check to the run's IP, the link needed its xsec_token, or your filters removed notes. RUN_SUMMARY lists every link that did not open and why the run stopped. Nothing is charged for rows that were not delivered.

Your feedback

Something missing or wrong? Open an issue on the Issues tab — we read every one.