WeChat Official Account Scraper → Markdown 公众号

Convert any WeChat Official Account (微信公众号) article into clean, RAG-ready Markdown plus structured metadata — title, author, account name, publish time, cover image and all image URLs. Built for Chinese content teams, AI researchers and data engineers feeding WeChat content into LLM / RAG pipelines.

Why this actor

• No official API exists for reading WeChat Official Account articles. This actor fills the gap.
• No API key, no WeChat account, no cookies needed for direct article URLs — article pages (mp.weixin.qq.com/s/...) are fetched with plain HTTP and parsed server-side.
• Markdown-native output: the article body (#js_content) is converted with Turndown, with WeChat's lazy-loaded data-src images rewritten to real URLs so no image is lost.
• Pay-per-event pricing: you only pay per article scraped. No start fee, no monthly minimum.

Reliability — read this

Mode	Reliability
Direct article URLs (articleUrls)	✅ Fully reliable, keyless, no proxy needed
Keyword discovery (searchKeywords)	⚠️ Best-effort. Uses Sogou WeChat search, which is captcha-prone. Without proxies it may work for a few queries, then get blocked. For production discovery use residential proxies (see proxyConfiguration). If discovery fails, the actor logs a warning and continues — direct URLs always work.

Input

Field	Type	Default	Description
articleUrls	string[]	[]	Direct WeChat article links (https://mp.weixin.qq.com/s/...). Fully reliable mode. No placeholder is prefilled — paste your own links.
searchKeywords	string[]	—	Optional keywords for article discovery via Sogou WeChat search. Best-effort, captcha-prone.
maxArticlesPerKeyword	integer	5	Max articles scraped per keyword.
includeImages	boolean	true	Keep  image markdown in the body. Image URLs are always collected in images regardless.
proxyConfiguration	object	—	Optional Apify proxy settings. Residential proxies recommended for keyword discovery.

At least one of articleUrls / searchKeywords is required — the actor fails fast with a clear error otherwise.

Example input

{
  "articleUrls": [
    "https://mp.weixin.qq.com/s/4gG0wmpjHS94fXEmAbfIXQ"
  ],
  "searchKeywords": ["人工智能"],
  "maxArticlesPerKeyword": 5,
  "includeImages": true
}

Output

One dataset item per article:

Field	Type	Description
source	string	"url" (direct link) or "search" (keyword discovery)
keyword	string | null	The keyword that found this article, or null for direct URLs
articleUrl	string	Canonical article URL
title	string	Article title
author	string | null	Article author(s)
accountName	string	Official Account (公众号) name
publishTime	string | null	ISO 8601 publish timestamp
coverUrl	string | null	Cover image URL
markdown	string	Full article body as clean Markdown
textLength	number	Plain-text length of the article body (chars)
imageCount	number	Number of images found in the body
images	string[]	Image URLs (capped at 50)
scrapedAt	string	ISO 8601 scrape timestamp

Example output item (markdown truncated)

{
  "source": "url",
  "keyword": null,
  "articleUrl": "https://mp.weixin.qq.com/s/4gG0wmpjHS94fXEmAbfIXQ",
  "title": "欧盟ESG法规解读系列之《零毁林法案》对中国新能源汽车企业的影响",
  "author": "苏畅 刘阳禾",
  "accountName": "金杜研究",
  "publishTime": "2025-04-07T10:51:51.000Z",
  "coverUrl": "https://mmbiz.qpic.cn/mmbiz_jpg/.../0?wx_fmt=jpeg",
  "markdown": "# 欧盟ESG法规解读系列...\n\n![](https://mmbiz.qpic.cn/...)\n\n...",
  "textLength": 5210,
  "imageCount": 6,
  "images": ["https://mmbiz.qpic.cn/..."],
  "scrapedAt": "2026-08-19T18:55:00.000Z"
}

Pricing

$5.00 per 1,000 articles — pay-per-event (article event), no start fee. You are charged only for articles successfully pushed to the dataset.

Usage tips

• Feed the markdown field straight into your chunking/embedding pipeline — it's already cleaned of WeChat's wrapper noise.
• WeChat image URLs (mmbiz.qpic.cn) are hotlink-protected for browsers but downloadable server-side; fetch them soon after scraping if you need to archive them.
• Sogou-discovered article URLs are signed and time-limited — scrape them promptly, or re-run discovery to get fresh links.
• For large keyword-discovery jobs, enable residential proxies to avoid Sogou captchas.

Limitations

• Deleted / account-restricted articles return an error page and are logged as failures (not charged).
• "Profile/list" pages (mp.weixin.qq.com/mp/profile_ext?...) are short-lived signed links and are not supported — use direct article URLs or keyword discovery.
• Keyword discovery depends on Sogou and can break or be captcha-blocked at any time; direct article URLs are the stable path.