WeChat Articles Scraper: search Official Account articles and get the full text

Search WeChat Official Account articles (公众号文章) by keyword without an account or a WeChat login, and get the complete article body in the same run: title, account, author, publish date, digest, cover, images and the full text, ready for analysis.

Built for Chinese-language corpora and LLM datasets, media and brand monitoring in China, competitor and industry research, and anyone who needs to read what Chinese companies, media and creators publish on WeChat, where most of that content lives.

What you need	What this Actor does	Price here
Articles for a keyword	title, account, date, digest, cover and link, ranked as on WeChat search	US$ 4.00 / 1,000
The full article	complete text, author, account id, all images	+ US$ 10.00 / 1,000
Articles you already have	send mp.weixin.qq.com links, get the full article	US$ 14.00 / 1,000

So a keyword search with the full text of every article costs US$ 14 per 1,000 articles. Rows that could not be delivered (a deleted article, a blocked page, a keyword with no results) come back with a stable errorCode and are never charged. Articles older than your date filter and duplicates across keywords are skipped and not charged either.

What you can ask for

You send	You get
queries: 新能源汽车, 跨境电商, AI agent	up to about 100 articles per keyword, in WeChat's relevance order
includeContent: false	only the search data, for a cheaper scan of what is being published
publishedAfter: 2026-01-01 or 30	only articles from that date on, or from the last 30 days
articleUrls: https://mp.weixin.qq.com/s/...	the full article for each link

Keywords work in Chinese and English, but WeChat is a Chinese-language platform: Chinese keywords return far more.

Output

One row per article.

{
  "type": "article",
  "searchQuery": "咖啡",
  "position": 3,
  "page": 1,
  "title": "咖啡自由后,我们离“瑰夏自由”还有多远?",
  "summary": "…",
  "accountName": "中国新闻周刊",
  "accountId": "gh_…",
  "author": "肖皖璇",
  "publishedAt": "2023-06-13T…Z",
  "digest": "…",
  "coverImage": "http://mmbiz.qpic.cn/…",
  "content": "full article text, one paragraph per line…",
  "contentLength": 3141,
  "images": ["http://mmbiz.qpic.cn/…"],
  "imageCount": 9,
  "articleUrl": "https://mp.weixin.qq.com/s?src=11&timestamp=…&signature=…",
  "scrapedAt": "2026-09-25T21:00:01Z",
  "errorCode": null,
  "error": null
}

The dataset has two views: Overview (title, account, author, date, size) and Full text.

Good to know

• About 100 articles per keyword. That is the most WeChat search shows without a login (10 pages). For more coverage, split a topic into several keywords: duplicates across them are removed and not charged.
• The article link is temporary. WeChat search hands out signed links that expire after a while. Save the text you need in the same run; that is why the full body comes included.
• Image posts are included. WeChat's photo posts (小绿书) come with their caption as content, every photo in images and postType: "image_post".
• No read or like counts. Those need a logged-in WeChat client, and this Actor never uses one.
• Deleted or restricted articles come back as a row with contentError and no charge for the text.

Errors you may see

errorCode	Meaning	Charged
NO_RESULTS	no article for this keyword in the chosen period	no
BLOCKED	the search page did not come back after several attempts; run again	no
NOT_REACHED	the run timeout arrived before this keyword	no
INVALID_URL	the link is not an mp.weixin.qq.com article	no
INVALID_INPUT	publishedAfter is not a date or a number of days	no
LINK_NOT_RESOLVED	the search listed the article but did not release its link; you still get title, account, date and digest	no

FAQ

Do I need a WeChat account or cookies? No. Nothing to log in to, nothing to paste.

How fast is it? About 10 articles every 10 to 20 seconds with the full text on, a bit faster with it off.

Can I use it from an AI agent? Yes, through the Apify MCP server or the Apify API, like any Actor.

What if a run is cut by its timeout? The Actor stops a few seconds before the limit and ends successfully with everything it already delivered, telling you in the status message how to get the rest.