Weibo 微博 Scraper

The Weibo Scraper extracts public data from Weibo (微博), China's largest microblogging platform. One actor, six modes: search posts by keyword, read user profiles and timelines, fetch posts by ID, collect comments and replies, or take the realtime hot-search list — no login.

中文说明

微博公开数据抓取工具。六种模式：按关键词搜索微博（综合或实时，可按日期范围）、获取用户资料、获取用户最新微博、按 ID 获取微博全文、获取评论与楼中楼回复、获取实时热搜榜。

每行数据都带 kind 标明类型：微博含正文 text_raw、发布时间 created_at、转发 reposts_count、评论 comments_count 与点赞 attitudes_count；用户资料含粉丝数 followers_count 与认证信息；热搜含排名 rank、热度 heat 与标签 label。

运行自带住宅代理，无需登录，也无需自行配置。结果支持导出 JSON、CSV 或 Excel。

What you get

One dataset, one row per record, with kind saying what each row is. Export as JSON, CSV or Excel.

{
  "kind": "post",
  "mid": "5345926956713830",
  "mblogid": "RjgLJsazA",
  "url": "https://weibo.com/2803301701/RjgLJsazA",
  "created_at": "2026-09-22T12:36:12+08:00",
  "text_raw": "【#南京通报幼儿园多名幼儿呕吐#】今天，南京江北新区联合调查组发布情况通报：……",
  "is_long_text": true,
  "source": "微博网页版",
  "user_id": "2803301701",
  "user_screen_name": "人民日报",
  "user_verified": true,
  "user_verified_reason": "《人民日报》法人微博",
  "user_followers_count": 157982893,
  "reposts_count": 17,
  "comments_count": 56,
  "attitudes_count": 125,
  "topics": ["南京通报幼儿园多名幼儿呕吐"],
  "pic_num": 1,
  "image_urls": ["https://wx3.sinaimg.cn/large/0033ImPzly1ihccc4ojdlj61uk0ycdy102.jpg"],
  "is_repost": false,
  "is_ad": false,
  "source_user": "2803301701",
  "scraped_at": "2026-09-22T05:01:32+00:00"
}

Every mode writes into the same dataset, and the Output tab has one view per kind:

• post — mid, mblogid, url, created_at, text_raw, is_long_text, the author (user_id, user_screen_name, user_verified, user_followers_count), reposts_count, comments_count, attitudes_count (likes), reads_count, region_name, topics, image_urls, video_url, video_play_count, the reposted post (retweeted_mid, retweeted_text_raw), is_ad
• profile — uid, screen_name, description, followers_count, friends_count (following), statuses_count, verified, verified_reason, domain, account_created_at, labels, avatar_url
• comment — comment_id, post_mid, is_reply, root_id, reply_to_id, floor_number, created_at, text_raw, like_count, reply_count, liked_by_author, the commenter (user_id, user_screen_name, user_verified)
• hot_search — rank, word, heat, label (新 / 热 / 沸), note, is_pinned, url

A post row says which target found it: source_keyword for a search, source_user for a timeline. A profile row carries source_input, the text you asked for. Authors and commenters ship their id, screen name, verification and counts only; location, gender and birthday appear on a profile only when the account is verified.

What it costs

Event	Price
Actor Start (apify-actor-start)	$0.00005
post (post)	$0.005
profile (profile)	$0.005
comment (comment)	$0.005
hot search entry (hot-search-entry)	$0.005

Input

Field	Type	Required	Default	What it does
mode	string	no	"hot_search"	What to collect: hot_search the realtime hot-search list (needs no other input); search posts by keyword; profile user profiles; timeline a user's recent posts; post specific posts with full text; comments comments and replies on specific posts. Each mode reads only its own fields — a field filled in for another mode is rejected, not ignored, so clear it when you switch. Default hot_search, the one mode that runs with nothing else filled in.
keywords	array	no	—	Search terms, one per entry — for example 咖啡 (coffee) or #新能源# for a topic. Each term returns up to max_items posts; a post matching two terms is returned once. Terms that differ only in surrounding spaces are searched, and billed, once. Required when mode is search; used ONLY by search — leave empty for every other mode.
users	array	no	—	Weibo users, one per entry: a numeric user id (2803301701), a screen name (人民日报), a custom domain (rmrb), or a profile URL such as https://weibo.com/u/2803301701 or https://m.weibo.cn/u/2803301701. The same user given in two forms is scraped, and billed, once. Required when mode is profile or timeline; used ONLY by those two — leave empty for every other mode.
post_ids	array	no	—	Weibo posts, one per entry: a numeric post id (mid, 16 digits), the short code after the user id in a weibo.com post URL (mblogid), or the post URL itself. The same post given in two forms is scraped, and billed, once. Required when mode is post or comments; used ONLY by those two — leave empty for every other mode.
max_items	integer	no	50	How many records to return per keyword, user or post, from 1 to 10000. Replies count toward a post's total like any comment. For hot_search it caps the whole run (the list holds about 50 entries). Default 50.
search_type	string	no	—	Which search results to read. top (综合, used when this is left empty) reads Weibo's relevance-ranked pages, which end after about 2,000 posts per keyword — often fewer. latest (实时) reads newest-first and keeps going back in time past that point, so it goes deeper; it also includes related posts that do not contain the literal keyword. A date range always searches the latest ordering, so do not combine top with dates. Used ONLY when mode is search — leave empty for every other mode.
start_date	string	no	—	Keep only posts published on or after this date, written as YYYY-MM-DD in Beijing time, for example 2026-09-01. Through the API a full timestamp such as 2026-09-01T08:00:00+08:00 also works. Leave empty for no lower bound. Setting a date searches the latest ordering, back through time, and drops posts outside the range. Used ONLY when mode is search — leave empty for every other mode.
end_date	string	no	—	Keep only posts published on or before this date (the whole day counts), written as YYYY-MM-DD in Beijing time, for example 2026-09-15. Leave empty for no upper bound. Must not be before start_date. Used ONLY when mode is search — leave empty for every other mode.
comment_sort	string	no	"hot"	Which comments come first: hot (most-liked first, the default) or time (newest first). Used ONLY when mode is comments; any other mode must keep the default.
include_replies	boolean	no	false	Also walk each comment's reply thread and return the replies as rows marked is_reply, linked to their comment by root_id. Replies count toward max_items. Default false. Used ONLY when mode is comments; any other mode must leave it off.
proxyUrls	array	no	—	Leave empty and the run uses the residential proxy this actor ships with, included in the price of the run. To route the run through your own account instead, add one gateway URL per entry, for example http://user:pass@host:port — works with DataImpulse, Bright Data, Oxylabs, Smartproxy or any provider that issues URLs. When set, only these URLs are used.

How to run it

1. Open the actor on Apify Console and pick a What to scrape mode — it decides which other fields matter. Each mode reads only its own fields; clear a field you filled for another mode.
2. Fill that mode's target: keywords for search, users for profile and timeline, posts for post and comments. Hot search needs nothing.
3. Run. Export JSON, CSV or Excel from the dataset.

{
  "mode": "search",
  "keywords": ["咖啡", "新能源"],
  "search_type": "latest",
  "start_date": "2026-09-01",
  "end_date": "2026-09-07",
  "max_items": 500
}

Minimal input:

{
  "mode": "hot_search"
}

Available as the MCP tool herus13--weibo-scraper on mcp.apify.com and through the Apify API; send the same JSON.

Use cases

• Brand and product monitoring — Search a brand name on the latest ordering every day and track how many posts, reposts and likes it draws, with the author's follower count to weigh reach.
• Trend tracking — Take the hot-search list on a schedule to see what China is talking about right now, with each entry's heat and its 新 / 热 / 沸 label.
• Public-opinion research — Collect every post for a keyword inside a date range, then the comments and replies under the most-discussed ones, as plain text ready for a classifier.
• Influencer and media vetting — Profiles with follower, following and post counts plus verification, next to the account's recent timeline.
• Campaign reporting — Fetch your own campaign posts by ID to record their repost, comment and like counts at a point in time.

FAQ

Do I need a Weibo account? No. Every mode runs anonymously — there is nothing to log in with and no cookies to manage.

How deep does a keyword search go? The default top ordering reads Weibo's relevance-ranked pages, which end after about 2,000 posts per keyword and often fewer. Choose latest, or give a date range, to keep going back through time past that point.

Why did my run stop with "does not use"? Each mode reads only its own fields, and a field filled in for another mode is refused rather than silently ignored — a date range left on a timeline would otherwise quietly return the newest posts. Clear the field, or switch back to the mode that uses it.

Why is a post's text cut short? Search and timeline rows carry Weibo's truncated text for long posts and set is_long_text. Run the post's mid through the post mode to get the full text.

Why do image links return 403? Weibo's image servers answer only requests that carry Referer: https://weibo.com/, and some links expire. Download images promptly and send that header.

How far back does a timeline go? About 990 posts, the most Weibo shows an anonymous visitor.

Is the data in Chinese? Yes — text, screen names and hot-search words come from Weibo as published, in Chinese.

Is a proxy included? A residential proxy is included in the price of the run. To use your own, paste gateway URLs into Your own proxy URLs. Scraping at volume? Your own DataImpulse account is cheaper per GB.

Is scraping Weibo legal? This actor reads only public posts and profiles, the ones any visitor sees, and never signs in. Personal details of ordinary users are deliberately left out. You are responsible for how you use the data — check Weibo's terms and your local rules.

Related actors

Building a wider social-listening pipeline? Pair this actor with:

• Bilibili Video Scraper — videos, creators, comments and danmaku from China's largest video community.
• TikTok Video Scraper — a creator's videos with play, like and share counts.
• Instagram Comment Scraper — comments and replies from public Instagram posts.
• YouTube Transcript Scraper — full transcripts of YouTube videos.