Weibo Scraper - Posts, Search, Comments & Hot Search

What does Weibo Scraper do?

Weibo Scraper extracts public data from Weibo (微博, Sina Weibo): posts by keyword with a date window, posts by link, an account's posts and profile, full comment threads with replies, likers, fans and followed accounts, the 热搜 hot-search boards, the 热门 hot-post feed, topic search and search suggestions. One dataset row per record, recordType says which. Built for brand and social-listening teams, China-market researchers, agencies vetting KOLs and data teams who need Weibo as clean JSON without a Weibo account.

Weibo scraper · 微博爬虫 · Sina Weibo scraper · Weibo API · Weibo hot search scraper · 微博热搜

• Keyword search that goes back in time: newest first, as far back as your date window, e.g. 1,000 posts over 3 days in one run (measured)
• Every post field: full text of long posts, likes, comments, reposts, reads, where it was posted from (发布于), topics, mentions, links, pictures, video stream and views, author stats
• Whole comment threads: 1,000 comments + replies of one post in 2 min 44 s, with each commenter's IP region (来自)
• Accounts: profile with followers, posts, verification, birthday, company, credit level; plus their newest posts and older video and picture posts
• Hot search: realtime 热搜 (~50 + pinned), 文娱 / 生活 / 社会 boards, the 热门 feed and 40 channels
• No login, cookies or Weibo API key: paste keywords, links, ids or account names
• Monitor mode: each scheduled run returns only new posts, comments and board changes (NEW / CHANGED / EXPIRED)
• Extras: sentiment, translation into 14 languages, contacts in bios, province rollup, compact rows for AI agents

What data can you get from Weibo?

📝 Post text (text)	🔗 Post link (postUrl, postBid)	📅 Posted at, Beijing time (createdAtIso)
❤️ Likes (likeCount)	💬 Comments (commentsCount)	🔁 Reposts (repostsCount)
📍 Posted from (region, 发布于)	#️⃣ Topics, @mentions, links	👀 Reads (readsCount)
🖼️ Pictures (images)	🎬 Video URL, length, views	📣 Promoted flag (isPromoted)
👤 Author, followers, verification	🪪 Profile: bio, birthday, company, credit	👥 Fans, following, likers
💬 Comment + IP region (ipRegion)	↩️ Replies (parentCommentId)	🔥 Hot-search rank + hot value
🏷️ Board label (新 / 热 / 沸) and category	🙂 Sentiment (sentiment)	🌐 Translation (translatedText)

On a 1,000-post run for 新能源汽车, 100% of posts had text, date, likes, comments, reposts and author follower count; 80.7% had the place they were posted from; 66% had topics; 55.7% had pictures and 25.6% a video.

How to scrape Weibo with Weibo Scraper

1. Create a free Apify account - $5 of monthly credit, no card needed.
2. Open Weibo Scraper and click Try for free.
3. Type keywords into Search keywords (or paste account or post links), and set Posted after for a date window.
4. Switch on the extras you want: Comments, Author profiles, Translate to, Hot-search boards.
5. Click Start. Rows stream into the dataset as they are found.
6. Open Storage and Export as JSON, CSV or Excel - or put the run on a schedule with Only new rows on.

⬇️ Input

Parameter	Type	Description
🔍 What to scrape		Fill any of these; every box takes links, ids or names.
searchKeywords	List	Keywords or #topics#, e.g. 咖啡, #新能源汽车#. Posts newest first, back through your date window.
profileUrls	List	Accounts: https://weibo.com/u/2803301701, 2803301701, 人民日报 or a custom address like rmrb.
postUrls	List	Posts: https://weibo.com/2803301701/RnAv0n3IZ, https://m.weibo.cn/detail/5356218505495865, a mid or a bid.
hotBoards	Multi-select	realtime 热搜 (~50 + pinned), entertainment 文娱, life 生活, social 社会 (50 each).
hotTimelineChannels	Multi-select	The 热门 hot-post feed and its channels (社会, 财经, 科技, 汽车, 美食, ... 40 in all).
searchUsers	List	Account search by keyword (Weibo shows the first 20 accounts).
searchTopics	List	Topic search: #topic# names with discussions and reads.
searchSuggestions	List	What Weibo suggests under the search box, with its count.
startUrls	List	Any mix of s.weibo.com searches, m.weibo.cn links, profile, post and t.cn links.
🎯 Filters (applied by Weibo)
searchSort	Select	realtime 实时 latest first (default), comprehensive 综合, hot 热门.
postedAfter / postedBefore	Text	Date window, Beijing time: 2026-09-01, 7 days, 24 hours, a day count.
✂️ Refine results (checked before a row is delivered)
originalOnly, verifiedOnly	Boolean	Drop reposts; keep V-verified authors only.
contentType	Select	all, pictures, videos, media, links, text.
adFilter	Select	all, organic_only (drop promoted), promo_only.
minLikes, minReposts, minComments, minFollowers	Integer	Minimums, 0 = none.
provinces, excludeKeywords, matchAllWords	List / Boolean	Posted-from places (北京, 广东, 海外), words to drop, every search word must appear.
➕ Extra data
includeComments, maxComments, commentSort	Boolean / Integer / Select	Comments of every delivered post; time = whole thread newest first (default), hot = hottest first. Default 20 per post.
includeReplies, maxRepliesPerComment	Boolean / Integer	Replies under each comment. Default 10.
commentsOnly	Boolean	Only the comments of the posts found (post rows not delivered or charged).
maxLikersPerPost	Integer	Accounts that liked each post, up to ~150.
maxFollowersPerUser, maxFollowingPerUser	Integer	First page of fans / followed accounts (up to 20).
includeAuthorProfiles	Boolean	One full profile row per post author.
fetchFullText	Boolean	Full text of long posts. Default on, free.
sentimentAnalysis, extractContacts, geoRollup	Boolean	Sentiment label + score; e-mail / phone / WeChat / QQ in bios; a GEO_ROLLUP record of posts per province. Free.
translateTo	Select	en, es, pt, fr, de, ru, ja, ko, vi, id, th, ar, tr, it - adds translatedText.
🔁 Monitoring
monitorMode, monitorStoreName, followUpComments	Boolean / Text / Boolean	Only rows earlier runs did not deliver; boards report NEW / CHANGED / EXPIRED; follow-up comments on the last 20 posts per keyword.
📦 Output
maxResults	Integer	All rows together; 0 = no total cap.
maxPostsPerSearch, maxPostsPerUser, maxHotTimelinePosts, maxTopicsPerBoard	Integer	Per keyword (default 100), per account (20), per channel (50), per board (60).
includeProfile, includeMediaPosts	Boolean	Profile row per account; older video and picture posts past the newest 10.
compact, includeRaw	Boolean	Core fields only (for AI agents); Weibo's own object as raw.
⚙️ Advanced
proxyConfiguration, residentialFallback, maxConcurrency	Proxy / Boolean / Integer	Apify datacenter proxy by default, then the container's own IP; your own proxy URLs work too.

Input JSON

{
  "searchKeywords": ["新能源汽车", "#小米汽车#"],
  "postedAfter": "7 days",
  "maxPostsPerSearch": 500,
  "includeComments": true,
  "maxComments": 50,
  "hotBoards": ["realtime"],
  "translateTo": "en"
}

Use the latest-first order with a date window

searchSort: "realtime" reads posts newest first and keeps going back in time until postedAfter or maxPostsPerSearch stops it. Do: {"searchKeywords": ["咖啡"], "postedAfter": "2026-10-01", "postedBefore": "2026-10-07"} · Don't: expect comprehensive to cover a whole week evenly; it follows Weibo's relevance ranking.

Search in Chinese for brands

Weibo indexes Chinese text, so 小米 finds many more posts than Xiaomi. Put both in searchKeywords; a post found by two keywords is delivered once.

Paste any Weibo link - it is routed for you

A profile link in the keyword box, a post link in the account box, an s.weibo.com search or a t.cn short link all go to the right job; a line that is not a Weibo link gets a free error row explaining why.

⬆️ Output

Each record is one dataset item. A post row (real, shortened):

{
  "recordType": "post",
  "postId": "5356218505495865",
  "postBid": "RnAv0n3IZ",
  "postUrl": "https://weibo.com/2803301701/RnAv0n3IZ",
  "text": "【夜读：#人生需要很多此刻就很好#】很多时候，我们活得不够洒脱 ...",
  "createdAtIso": "2026-10-20T22:11:08+08:00",
  "likeCount": 1775,
  "commentsCount": 1293,
  "repostsCount": 1295,
  "region": null,
  "topics": ["人生需要很多此刻就很好", "幸福就在当下的每一分每一秒里"],
  "postType": "text",
  "authorId": "2803301701",
  "authorName": "人民日报",
  "authorFollowers": 158103656,
  "authorVerified": true,
  "sourceType": "url"
}

A comment row: {"recordType": "comment", "commentId": "5356353963165627", "postId": "5356218505495865", "text": "[心]", "createdAtIso": "2026-10-21T07:09:24+08:00", "likeCount": 0, "replyCount": 0, "floorNumber": 1424, "ipRegion": "辽宁", "isReply": false}. A trend row: {"recordType": "trend", "board": "realtime", "rank": 1, "word": "男子捡到38000元留了道题给失主", "hotValue": 1169412, "label": "热"}.

Fields

Fill measured on real runs: posts = 1,020 rows (a 1,000-post search + the prefill), comments = 1,146 rows, accounts = 115 rows of all six kinds, trends = 69 rows.

Field	Type	What it is	Fill
postId, postBid, postUrl, postUrlMobile	string	Ids and links of the post	100%
text, textLength, isLongText, isTruncated	string / number / boolean	Plain text (full text of long posts)	100%
createdAtIso, createdAtTimestamp, createdAt	string / number	When it was posted (Beijing time)	100%
likeCount (= attitudesCount), commentsCount, repostsCount	number	Engagement	100%
readsCount	number	Reads (posts read by link)	posts by link
region	string	Where Weibo says it was posted from (发布于)	80.7%
topics, mentions, links	array	#topics#, @mentions, outside links	66% / 4% / 31%
images, picCount	array / number	Pictures	55.7%
videoUrl, videoStreams, videoDuration, videoViews, videoTitle, videoCover	string / object / number	Video posts	25.6%
postType, isRepost, repostOf, isPromoted, cardType, cardTitle, cardUrl, source, editedCount	mixed	Kind of post, repost source, promoted flag, attached card, client, edits	100% (cards 42%)
authorId, authorName, authorFollowers, authorFollowing, authorPostsCount, authorVerified, authorVerifiedReason, authorDescription, authorGender	mixed	Author	100% (bio 72%, reason 59%)
commentId, postId, text, createdAtIso, likeCount, replyCount, floorNumber	mixed	Comment	100%
ipRegion	string	Where the comment came from (来自)	99.6%
isReply, parentCommentId, replyToUserName, isByPostAuthor, likedByPostAuthor, picUrls	mixed	Thread links, author flags, pictures	as present
userId, screenName, followersCount, friendsCount, verified, relation	mixed	Account rows (profile, author, fan, following, liker, search)	100%
statusesCount, description, gender, verifiedReason	mixed	Fuller account fields (likers carry fewer)	30-36% of all account rows
birthday, company, accountCreatedAt, sunshineCredit, location, labels	string / array	Profile and author rows only	on profile rows
board, rank, word, hotValue, label, category, isPromoted	mixed	Hot-search rows	word 100%, hotValue 85-98%
discussions, reads	number	Topic-search rows	on topic rows
change, rankDelta, hotValueDelta, previousRank, peakRank, minutesOnBoard	mixed	Monitor mode on boards	monitor runs
sentiment, translatedText, contacts	object / string	When those extras are on	when on

How can I use Weibo data?

• 📣 Brand and social-listening teams - track a brand name every day on the latest-first order, with likes, reposts and where each post came from
• 🔥 Trend and news desks - record the 热搜 boards every 10 minutes with monitor mode and get only what moved
• 🧭 China-market researchers - collect every post for a topic inside a date range, then the comment threads under the biggest ones
• 🤝 KOL and agency teams - profiles with followers, posts, verification and credit level, plus their recent posts and comment sentiment
• 🌐 Global teams - translate posts, comments and board topics into English as they are collected
• 🤖 AI and data teams - Chinese social text with dates, regions and sentiment, compact rows for LLM pipelines

How much does it cost to scrape Weibo?

Weibo Scraper is pay-per-result with no start fee:

Row type	Price per 1,000
Post	$0.81
Account (profile, author, fan, following, liker, account search)	$0.81
Comment or reply	$0.45
Hot-search topic, topic-search row or suggestion	$0.90
Translation (per row that carries translatedText)	$2.70

Example	Cost
100 posts for one keyword	$0.081
1,000 posts in a date window	$0.81
1,000 comments of one post	$0.45
The realtime hot-search board (~50 topics)	~$0.046
100 posts translated into English	$0.351

Full text of long posts, sentiment, contacts in bios and the province rollup are free. Filtered posts and rows not delivered are not charged. Measured speed on the default 256 MB: 100 posts in 28 s, 1,000 posts in 7 min 47 s, 1,002 comments in 2 min 44 s.

Integrations, API, scheduling & MCP

• Schedules: run Weibo Scraper hourly, daily or every 10 minutes from Schedules; with Only new rows each run returns just what is new.
• Webhooks: call a URL when a run finishes to push rows into your CRM, Slack or a sheet.
• No-code: Make, Zapier, n8n, Google Sheets, Airtable and Slack through Apify integrations.
• API: start runs and read datasets from Node.js or Python with apify-client - code on the API tab, reference at docs.apify.com/api/v2.
• MCP / AI agents: connect Claude, Cursor or any MCP client through the MCP tab and ask for Weibo posts or hot search in plain language.

More TopApi scrapers

🎵 Douyin Scraper - 抖音 videos, creators, comments, hot search	🏭 1688.com Wholesale Scraper - Chinese suppliers, prices, MOQ	🎬 TikTok Scraper - videos, hashtags, profiles
📸 Instagram Scraper - posts, reels, hashtags, profiles	🐦 Twitter Profile Scraper - X followers, bio, verified	🗽 Truth Social Scraper - posts, replies, profiles
🔎 Facebook Search Posts Scraper - keyword, date, location	🧭 All TopApi actors

Limits & empty runs

• An account's posts: Weibo shows a signed-out visitor an account's newest 10 posts; Weibo Scraper adds the account's older video and picture posts from its video and album tabs (3 accounts x 50 posts measured). Older text-only posts are shown to signed-in users only.
• Lists Weibo shows signed-out visitors only in part: account search 20 per keyword, fans and following the first page (some large accounts hide their fans), likers about 150 per post. Repost lists are not offered: Weibo shows them to signed-in users only.
• Comments: newest-first reads a whole thread (1,277 of 1,288 measured); hottest-first stops after about 250-300.
• Speed: about 2-3 posts a second for a single keyword on 256 MB; several keywords run side by side.
• Empty dataset: no post matched your filters or date window, the account or post does not exist (a free note in the status message and run summary), or nothing is new in monitor mode. Nothing is charged for any of these.

FAQ

How many posts can Weibo Scraper get for one keyword?

As many as Weibo has in your date window: it reads newest first and steps back in time until postedAfter or maxPostsPerSearch stops it. A 3-day window for 新能源汽车 returned 1,000 posts in one run.

Can Weibo Scraper get Weibo comments and replies?

Yes. Switch on Comments (and Replies). Newest-first reads a whole thread; each comment carries likes, reply count, floor number and the commenter's IP region.

Can I track Weibo hot search (微博热搜) with Weibo Scraper?

Yes. Pick the realtime, 文娱, 生活 or 社会 board. With Only new rows on, each scheduled run returns topics that are NEW, CHANGED (moved rank) or EXPIRED, with rank and hot-value changes.

Can I run Weibo Scraper on a schedule?

Yes. Save your input as a task, open Schedules, pick the cadence and switch on Only new rows. A run that finds nothing new delivers nothing and charges nothing.

Can I use Weibo Scraper with the Apify API?

Yes. Use apify-client for Node.js or Python, or call the REST API directly; the API tab has ready-to-paste code.

Can I use Weibo Scraper through an MCP server?

Yes. Open the MCP tab, connect your MCP client and ask in plain language. Switch on Compact rows to keep answers short.

Do I need a Weibo account, cookies or an API key for Weibo Scraper?

No. Weibo Scraper reads what Weibo shows any signed-out visitor; you pass keywords, links or account names and nothing else.

Can Weibo Scraper translate Weibo posts into English?

Yes. Set Translate to and each post, comment and board topic gets a translatedText field next to the original.

Why did my Weibo Scraper run return no rows?

Usually a filter or date window removed everything (the status message names the filter and the count), the account or post id does not exist, or monitor mode found nothing new since the last run.

Is it legal to scrape Weibo?

Weibo Scraper reads only public posts and profiles, the ones any visitor sees, and never signs in. Results can contain personal data; make sure you have a legitimate purpose for how you use them.

Your feedback

Found a bug or missing a field? Open an issue on the Issues tab - we read every one.