WeChat Official Account Scraper - Articles, Accounts & Full Text

Search WeChat Official Account (微信公众号) articles by keyword and get clean rows with English translations, exact publish times, working article links and optional full article text.

WeChat Official Accounts are where Chinese brands, state media, industry analysts and local competitors actually publish. There is no public WeChat search API, the in-app search is mobile-only and unscriptable, and the article links you find elsewhere are Sogou redirects that expire or bounce to a captcha. So most teams end up pasting Chinese text into a translator one article at a time, or paying per operation for a tool that caps a run at 50 rows. This Actor takes a list of keywords and hands back every article the source will serve, already translated, already timestamped, already deduplicated.

✅ What you get / ❌ what this isn't

✅ What you get	❌ What this isn't
Up to ~89 articles per keyword, as many keywords per run as you like	Not capped at 50 rows per run
English title and summary next to the Chinese original	Not a Chinese-only dump you still have to translate
Exact publish time as an ISO timestamp, from the page's own epoch	Not a vague "3 天前" relative string
A resolved mp.weixin.qq.com link that opens	Not the raw Sogou /link? redirect that 302s to a captcha
A stable articleId that is identical across runs	Not a row you have to fuzzy-match on title to dedup
Full article body text, author byline and account IDs	Not a truncated search snippet you have to enrich yourself
Priced per article returned, no start fee	Not a per-run fee charged before any row exists
Paste article links you already have and skip search	Not keyword-only
Only the two things the source actually does	Not a list of 15 "operations" where several return nothing

🔎 Why use this Actor

• No per-run ceiling. The source itself stops at about 89 articles per keyword. This Actor walks every page up to that ceiling and accepts any number of keywords in one run, so a 10-keyword run can return hundreds of articles.
• English alongside the original. titleEnglish and summaryEnglish sit next to title and summary, so a non-Chinese-reading analyst can scan the dataset directly. Turn it off with translateToEnglish if you want the raw Chinese only.
• Exact publish times. WeChat renders publish dates through a JavaScript call, so scrapers that read the visible text get a relative string. This one pulls the Unix epoch out of that call and emits both publishedAt (ISO) and publishedTimestamp.
• Links that actually open. Every row's Sogou redirect is resolved through the hop page into a real mp.weixin.qq.com URL. That hop is rate-limited and needs the search session's cookies, which the Actor carries for you.
• Dedup that survives across runs. articleId is the source's own document fingerprint, so re-running a keyword next week and filtering on it gives you only what is new.
• Full text by default. Each row carries contentText, author, accountBiz, messageId and messageIndex. Turn fetchFullContent off for a faster, cheaper run that returns search snippets only.
• Bring your own links. Drop mp.weixin.qq.com URLs into articleUrls and they are scraped directly, without going through search.
• Pay per article, nothing to start. No actor-start fee, so an empty run is free. Snippet-only rows are charged at a lower rate than full-text rows.

👥 Who it's for

China market researchers, brand and competitor monitoring teams, consultants tracking Chinese industry commentary, and anyone building a dataset of Chinese-language coverage.

• Track what Chinese official accounts publish about your brand, a competitor or a category.
• Build a translated reading list of Chinese industry commentary without reading Chinese.
• Feed contentText into your own summarizer, classifier or RAG index.
• Re-run weekly on the same keywords and keep only rows whose articleId you have not seen.
• Pull publisher and byline data to find which accounts dominate a topic.

Example tasks

Ready-made runs you can open and start without configuring anything:

• WeChat AI News Monitor - 人工智能 coverage
• WeChat New Energy Vehicle Coverage - 新能源汽车 coverage
• WeChat Health and Wellness Content - 大健康 coverage
• WeChat Livestream E-Commerce News - 直播电商 coverage
• WeChat Cross-Border E-Commerce Articles - 跨境电商 coverage

⚙️ How to search WeChat articles

1. Open the Actor and put one or more keywords into Search queries. Chinese keywords return the most results; English brand names work when written in Latin script. Or paste article links straight into Direct article URLs.
2. Set Maximum articles per query. The source tops out near 89, so anything above that simply returns everything available.
3. Leave Translate titles and summaries to English on unless you want raw Chinese.
4. Leave Fetch full article text on to get article bodies. Turn it off for a faster and cheaper run that returns only titles, summaries, accounts and links.
5. Click Start. Rows stream into the dataset as each page is parsed, so you can watch results arrive and export any time from the Storage tab as JSON, CSV or Excel.

Get only new articles each week

Save the run as a Task and schedule it. Each row carries articleId, the source's own stable fingerprint for that article, plus publishedAt. Filter incoming rows against the articleId values you already stored and you get a clean weekly delta with no duplicates, without re-reading anything.

📥 Input

{
  "searchQueries": ["人工智能", "新能源汽车"],
  "articleUrls": [],
  "maxArticlesPerQuery": 20,
  "fetchFullContent": true,
  "translateToEnglish": true
}

• searchQueries - list of keywords to search. Each returns up to about 89 articles. Required unless you supply articleUrls.
• articleUrls - optional mp.weixin.qq.com links you already have. These are scraped directly and skip search entirely.
• maxArticlesPerQuery - how many articles to keep per keyword. Default 20, maximum 100, though the source ceiling of ~89 applies first.
• fetchFullContent - open each article for the full body, author byline and account IDs. Default true. Turning it off returns search snippets only, at the lower price.
• translateToEnglish - add English title and summary. Default true.

📤 Output

titleEnglish	accountName	publishedAt	articleUrl
"Can't let China catch up": US AI giants turn hawkish	观察者网	2026-09-20T09:56:57.000Z	https://mp.weixin.qq.com/s?src=11&...
The new energy vehicle price war enters round three	财经十一人	2026-09-19T23:11:04.000Z	https://mp.weixin.qq.com/s?src=11&...
Three things the chip export rules changed overnight	半导体行业观察	2026-09-19T14:02:38.000Z	https://mp.weixin.qq.com/s?src=11&...

Every row also carries query (the keyword that found it), articleId (stable dedup key), title and summary in the original Chinese, summaryEnglish, publishedTimestamp (Unix epoch), sogouUrl (the unresolved source link, kept for traceability), thumbnailUrl (unwrapped to the original image host rather than the CDN wrapper that blocks hotlinking) and scrapedAt. With fetchFullContent on, rows add contentText, contentLength, author, accountNickname, accountBiz, messageId and messageIndex - the last three being WeChat's own identifiers for the publishing account and the specific message.

💵 How much does it cost?

You pay per article returned, with no actor-start fee, so a run that finds nothing costs you nothing. There are two rates: a lower one for an article returned from search (title, summary, account, timestamp, link) and a higher one for an article whose full text was fetched. That means fetchFullContent: false is genuinely cheaper, not just faster. A typical keyword yields 20-89 articles, so a three-keyword run lands somewhere around 60-260 rows.

🔁 Run it on the Apify platform

Schedule it daily or weekly from the Apify Console, or call it from the API and let results land straight in your warehouse. Results export as JSON, CSV, Excel or XML, and the dataset can be pushed onward through webhooks or the Make, Zapier and n8n integrations, which makes it straightforward to drop new Chinese-language coverage into Slack, Notion or a BI table as it appears.

⚠️ Limits and caveats

• About 89 articles per keyword is a hard source ceiling. The index stops serving after page 10 regardless of how many results it claims to have. Use more, narrower keywords to go deeper rather than raising maxArticlesPerQuery.
• There is no date filter and this Actor does not fake one. The source accepts a date parameter and then returns zero rows for every value of it, which would look exactly like a genuine no-match. Sort or filter on publishedAt after the fact instead.
• Results are relevance-ordered, not newest-first. The source ignores its own sort parameter. Sort on publishedAt yourself if you need chronological order.
• Account search is not offered. Looking up an official account by name is soft-blocked at the source: it returns a valid page with zero accounts even for famous publishers. Keyword article search is the surface that genuinely works.
• articleUrl is signed and time-limited. WeChat does not expose the sn token on this path, so a permanent canonical link cannot be built. The link works when fetched and for a while after; articleId, accountBiz and messageId are the durable identifiers.
• Link resolution is rate-limited even though search is not. On long runs some rows may arrive with articleUrl empty. The article itself is still in the row, and the run summary reports how many links could not be resolved.
• Coverage is the public index, not all of WeChat. Articles from accounts that block indexing, and subscriber-only posts, do not appear.
• Translation is machine translation. titleEnglish and summaryEnglish are good enough to triage and search on, not to quote. The Chinese original is always kept alongside.

🧩 Related Actors

• 1688 Product Scraper - pair keyword coverage with the wholesale prices, MOQs and suppliers behind the products being discussed.
• Weibo Scraper - the open, conversational half of Chinese social. WeChat carries the long-form publishing, Weibo the reaction.
• RedNote (Xiaohongshu) & Douyin Scraper - consumer product trends to sit alongside WeChat's editorial coverage.
• Shopee Scraper - take a category you found trending in WeChat coverage through to live marketplace listings.
• TikTok Hashtag Trends Scraper - the same trend question on the Western side of the same company's ecosystem.

❓ FAQ

Do I need a WeChat account or any login? No. This reads the public article index. There are no credential fields.

Am I charged for keywords that return nothing? No. Charging is per article returned, and there is no start fee, so an empty run is free.

Can I search in English? Yes, but expect fewer results. Official accounts publish in Chinese, so Chinese keywords return far more. English brand and company names do work when written in Latin script.

Can I get articles from one specific official account? Not directly. Account lookup is blocked at the source. In practice, searching the account's name as a keyword surfaces much of its output, and you can filter rows on accountName.

How do I avoid duplicates when I run this on a schedule? Keep the articleId values you have already seen and drop incoming rows that match. It is the source's own fingerprint for an article and does not change between runs.

Why is articleUrl empty on a few rows? The link-resolution step is rate-limited separately from search. Those rows still carry the title, summary, account and timestamp; the run summary reports the count.

🛠️ Support

Something wrong or missing? Open an Actor issue with the run URL, the input you used and what you expected. That is the fastest route to a fix.

⭐ Rate this Actor

If this saved you time, please leave a rating on the Reviews tab. Review count is the main trust signal buyers have on the Store, and reviews are what decide which Actors get improved next. If something is broken, please open an issue first so it can be fixed.