Weibo Search Scraper — Posts by Keyword · No Login

Search Weibo (Sina Weibo), China's largest social media platform, by keyword and get structured post data back: author, full text, timestamp, repost/comment/like counts and the post URL. No login, no cookies, no app token.

Why use this scraper

Weibo carries real-time Chinese consumer and public opinion at a scale no other platform matches for that market, but it has no public search API for third parties, and its web app renders results behind JavaScript that changes format without notice. Brands tracking how a product launch is received in China, researchers studying Chinese public discourse, and competitive intelligence teams watching a competitor's mentions all need a way to pull Weibo search results programmatically, in a language-agnostic structured format, without maintaining their own scraper against a site that reworks its front-end regularly.

This actor takes one or more keywords and returns the matching posts as clean JSON: who posted it, the full post text, when it went up, and its engagement numbers (reposts, comments, likes). That engagement data is what turns a keyword search into a signal: a spike in attitudesCount and commentsCount around a specific post is a far stronger indicator of a trending topic than the keyword match alone.

Overview

Input is a single keyword or a list of keywords; output is one JSON record per matching post, up to maxItems per keyword. Results are ordered the same way Weibo's own search returns them, so the first results are Weibo's own relevance/recency ranking rather than a random sample. A typical run for one keyword and 20 posts completes in well under a minute.

Supported inputs

• Single keyword — keyword, a plain search term such as a brand name, product name or hashtag topic.
• Multiple keywords — keywords, an array of terms searched independently in the same run, with results split per keyword so you can tell which search produced which post.
• Result cap per keyword — maxItems (1 to 200) controls how many posts are pulled for each keyword before the actor stops paginating.

Use cases

• Brand monitoring teams — track how a product name or campaign hashtag is performing on Weibo day to day, without a Weibo account.
• Market researchers — pull consumer sentiment and discussion volume around a category (e.g. a phone model) for a China market report.
• Competitive intelligence — watch a competitor's brand or executive name for spikes in mentions tied to news events or launches.
• Academic and media researchers — collect a structured sample of Chinese-language public posts on a topic for discourse analysis.
• Trend spotting — compare repostsCount and attitudesCount across keywords to see which topic is actually gaining traction versus just being mentioned.
• Crisis and PR monitoring — catch an early spike in negative-sentiment posts mentioning a brand or product before it escalates into a full news cycle.

How it works

The actor queries Weibo's own public search interface for the keyword and reads the same structured post data Weibo's mobile site renders for the results feed, so authors, counts and timestamps come back as clean fields instead of scraped HTML text. Multiple keywords are processed independently in the same run, each paginated up to its own maxItems.

Input configuration

Single keyword:

{
  "keyword": "iphone",
  "maxItems": 50
}

Multiple keywords in one run:

{
  "keywords": ["iphone", "小米18", "华为Mate"],
  "maxItems": 30
}

Small sample for a quick check:

{
  "keyword": "iPhone Air",
  "maxItems": 10
}

Output sample

Real output from a live search for "iphone":

[
  {
    "keyword": "iphone",
    "authorName": "老师好我叫何同学",
    "authorId": 6529876887,
    "text": "iPhone 18 Pro系列的首发体验视频来了！希望大家喜欢！#何同学##iPhone18上手评测##苹果发布会##苹果发布会有何亮点# 老师好我叫何同学的微博视频",
    "createdAt": "Wed Sep 16 20:29:16 +0800 2026",
    "repostsCount": 551,
    "commentsCount": 322,
    "attitudesCount": 4399,
    "source": "微博视频号",
    "url": "https://m.weibo.cn/detail/5343871682741457",
    "scrapedAt": "2026-10-10T09:17:25.432Z"
  },
  {
    "keyword": "iphone",
    "authorName": "钟文泽",
    "authorId": 1906286443,
    "text": "我问我妈，今年要不要给她换一台 iPhone Duo...",
    "createdAt": "Sat Oct 10 15:09:50 +0800 2026",
    "repostsCount": 11,
    "commentsCount": 209,
    "attitudesCount": 846,
    "source": "微博VLOG博主 微博2019十大影响力数码大V",
    "url": "https://m.weibo.cn/detail/5352488603749356",
    "scrapedAt": "2026-10-10T09:17:25.432Z"
  }
]

Key output fields

Field	Type	Description
keyword	string	The search term that matched this post (useful when running multiple keywords)
authorName	string	Display name of the posting account
authorId	integer	Weibo's internal numeric ID for the author
text	string	Full post text, including hashtags, as written
createdAt	string	Original post timestamp, as Weibo formats it
repostsCount	integer	Number of reposts at scrape time
commentsCount	integer	Number of comments at scrape time
attitudesCount	integer	Number of likes ("attitudes") at scrape time
source	string	Client or account label Weibo shows under the post (device, verified status label, etc.)
url	string	Direct link to the post
scrapedAt	string	ISO timestamp the actor collected this record

Pricing

Pay-per-event, charged only for posts actually returned:

Event	Price
Post scraped	$0.005 per post

1,000 posts cost $5.00. There is no separate charge for the search request itself, only for posts returned.

What makes this richer than the competition

Weibo scraping is a crowded category on the Store. sian.agency/weibo-scraper is the largest at 1,863 users in the last 30 days, and zhorex/weibo-scraper ("Hot Search, Posts, Comments & Creator Feeds") follows with 427; both cover a broader surface (hot search boards, full comment threads, creator feed crawling) and are built for teams that need the whole Weibo surface area. This actor is deliberately narrower: pure keyword search, returning the engagement numbers needed to judge whether a topic is actually trending, with no login step and no need to configure comment-depth or feed-crawl options. For a pipeline that only needs "what is Weibo saying about X right now, and how much traction does it have," this is a faster, simpler integration than a full-surface scraper.

Notes & limitations

• This actor scrapes Weibo's public search results only; it does not log in, so it cannot reach private accounts or posts restricted to followers.
• Engagement counts (repostsCount, commentsCount, attitudesCount) are a snapshot at scrape time; Weibo posts keep accumulating engagement after the run, so re-running later on the same post will show higher numbers.
• Result ordering follows Weibo's own search ranking (a mix of relevance and recency), not purely chronological order.
• Very broad single-character or extremely common keywords can return a wide mix of unrelated posts, since the match is on Weibo's own search index, not a strict phrase filter.
• Mixed Chinese/English keywords are supported, since the actor passes the term through to Weibo's search as-is.
• text is returned exactly as written by the author, including embedded hashtags (#topic#) and any trailing "全文" ("full text") marker Weibo itself shows for posts it has truncated in the search preview.
• There is no built-in date-range filter; if you need posts from a specific window, filter the output by createdAt after the run, or narrow the keyword to a more specific phrase that naturally limits results to a recent event.

FAQ

Do I need a Weibo account or API key? No. The actor reads Weibo's public search results directly; no login or app credentials are required.

Can I search in English or does it have to be Chinese? Both work. Weibo's own search handles English brand/product names (like "iphone") as well as Chinese terms.

How many posts can I pull per keyword? Up to 200 per keyword in a single run via maxItems. For more than 200, run the actor again or schedule repeated runs.

Does it return comments on the posts, or just the posts themselves? Just the posts matching the keyword search, with their engagement counts. It does not expand into comment threads.

Can I track a hashtag over time? Yes. Scheduling this actor daily against the same keyword and comparing attitudesCount/commentsCount growth is a common way to track a hashtag's momentum.

What does source mean in the output? It is the label Weibo itself displays under a post, typically the posting client or an account badge (e.g. a verified-influencer label); it comes straight from Weibo, not from this actor.

For AI Agents & LLM Apps

This actor is available over the Apify MCP server (mcp.apify.com) for direct use from Claude, Cursor or any MCP-compatible agent, and over the plain REST API for scripted pipelines.

Synchronous call that returns dataset items directly, no polling required:

curl "https://api.apify.com/v2/acts/gio21~weibo-search-scraper/run-sync-get-dataset-items?token=<APIFY_TOKEN>" \
  -X POST \
  -H "Content-Type: application/json" \
  -d '{"keyword": "iphone", "maxItems": 20}'

Disclaimer

This actor is not affiliated with, endorsed by, or sponsored by Sina Weibo Corp. "Weibo" is a trademark of its respective owner. The actor only collects data that is publicly visible on Weibo's own search results at the time of the run.

SEO Keywords

Weibo scraper, Weibo search API, Sina Weibo data, Weibo keyword monitoring, Chinese social media scraper, Weibo post scraper.