WeChat Official Account Articles Scraper

Collect the public article history of the WeChat Official Accounts (微信公众号) you follow: every article in the account's public albums (合集), newest first, plus the account's articles in Sogou's public WeChat index, each with the full text, author, publish time, images and account details. Turn on the checkpoint and scheduled runs return only articles you have not received yet.

No WeChat login, no cookies, no app and no third-party data API.

What does WeChat Official Account Articles Scraper do?

You give it accounts. It gives you their articles.

• 📚 Account history from albums. Reads every public album (合集) of the account through WeChat's own album pages, newest first, with permanent article links. Long albums are followed page by page.
• 🔎 Sogou-indexed articles. Searches Sogou's public WeChat index for the account name and keeps only articles published by that exact account. This also finds articles that are not in any album.
• 📄 Full article text. Reads each article page for the body text (with paragraph breaks), optional body HTML, every in-article image, author, publish location (发表于), cover, digest and the account's gh_ id and avatar.
• 🔁 Only new articles. A checkpoint per account remembers what was delivered. The next run skips those articles and, when there is room, continues further back into album history.
• 🔗 Article links. Paste mp.weixin.qq.com links to read single articles in full.

Who is it for?

• China media-monitoring teams archiving new articles from a fixed list of brand, media and government accounts every day.
• Market and competitor researchers reading a competitor's or a category leader's recent publishing in full text.
• Investment and industry analysts following research and newsletter accounts (行研, 周报) and feeding them into a reading pipeline.
• PR and communications teams keeping a dated record of what an account published and when.
• Data and AI teams building Chinese-language corpora or RAG indexes from known publishers, with stable article keys for deduplication.

Why use this WeChat article history scraper?

• History, not just search. Most WeChat tools are keyword search. This Actor starts from the account and walks its albums like a publishing archive.
• Checkpointed monitoring. onlyNew keeps one small record per account in your own key-value store. Daily schedules return only new posts; articles delivered before are skipped and cost nothing.
• Normalized article bodies. Clean text with paragraph breaks, optional cleaned HTML with real image URLs, and a character count, ready for search indexes and LLMs.
• Permanent identity. Every article whose page or album entry is read carries articleKey (__biz_mid_idx), which stays the same across runs even when a link is temporary. The one exception is listed under Tips and limits.
• Honest coverage fields. source tells you whether a row came from an album, the Sogou index or your link; contentStatus tells you whether the full text was read.

What data can you extract from WeChat Official Accounts?

Field	Description
title	Article title
publishedAt / publishedTimestamp	Publish time (ISO 8601 UTC / Unix seconds)
accountName	Official Account display name
accountId	Account gh_ id
accountBiz	Account __biz id
accountAvatarUrl	Account avatar
author	Signed author (byline)
digest	Summary under the title
contentText	Full body text with paragraph breaks
contentHtml	Cleaned body HTML (optional)
imageUrls	Images in the article body
wordCount	Non-space characters in the body
coverImageUrl	Cover image
publishLocation	发表于 province or country
albumId / albumTitle	Album (合集) of the article
articleUrl	Article link
articleUrlIsPermanent	Whether the link keeps working
articleKey, messageId, indexInMessage	Permanent article identity (null only for Sogou-only rows whose page was not read: full text off, or the page was blocked or unreadable)
source	album, sogou or article_url
contentStatus	ok, not_requested, blocked or unavailable
inputAccount, collectedAt	Which input produced the row, and when

Read counts, likes, comments and other figures that only the logged-in WeChat app shows are not included.

How much does it cost to scrape WeChat Official Account articles?

This Actor uses pay-per-event pricing. You pay a small start fee per run and then per saved article. Prices per article fall on higher Apify plans:

Event	Free	Bronze	Silver	Gold and above
Run start (once per run)	$0.005	$0.005	$0.005	$0.005
Article with full text	$0.00552	$0.0048	$0.003744	$0.00288
Article metadata (no full text)	$0.003105	$0.0027	$0.002106	$0.00162

"Article with full text" covers one saved article with its full text, author, images and account details. "Article metadata" covers one saved article without full text: title, date, link and cover, when full text is off or the page could not be read.

Examples on the Free plan:

• 20 articles with full text from one account: $0.005 + 20 × $0.00552 = $0.115
• 100 articles with full text from five accounts: $0.005 + 100 × $0.00552 = $0.557
• 500 album entries without full text: $0.005 + 500 × $0.003105 = $1.558

Duplicates, articles outside your date filter, articles skipped by the checkpoint and runs that stop on invalid input are free. Platform usage is included in these prices. Set a maximum cost per run in the run options to cap spending.

How to scrape a WeChat Official Account's articles

1. Open the Actor and add accounts under Official Accounts: the exact display name (for example 刀法研究所), a link to any one of its articles, or a link to one of its albums.
2. Set Max articles per account (default 20) and, if you only want recent posts, Published after (for example 30 days).
3. Keep Include full text on for the article body, or turn it off for a cheaper list of titles, dates and links.
4. For daily monitoring, turn on Only new articles and save the input as a Task with a schedule.
5. Click Start, then export the dataset as JSON, CSV, Excel or through the API.

Input parameters

Parameter	Type	Default	Description
accounts	array of strings	["刀法研究所"] when the input is empty	Display names, article links or album links, one account each
articleUrls	array of strings	–	Single article links. publishedAfter and includeContent apply to them; onlyNew does not
maxArticlesPerAccount	integer	20	Most articles per account per run, newest first (1–2000)
publishedAfter	string	–	YYYY-MM-DD (China time) or a period such as 7 days, 4 weeks, 6 months
includeContent	boolean	true	Read each article page for the full text
includeHtml	boolean	false	Also save cleaned body HTML
onlyNew	boolean	false	Skip articles delivered before; continue deeper on later runs
stateStoreName	string	wechat-account-history	Key-value store for the checkpoints
sogouPages	integer	3	Sogou index pages (10 results each) per account name; 0 turns it off
maxAlbumsPerAccount	integer	10	Most albums read per account

Input examples

The latest 20 articles of an account, with full text:

{ "accounts": ["刀法研究所"], "maxArticlesPerAccount": 20 }

Everything an account published in its albums in the last month, named by an article link:

{
  "accounts": ["https://mp.weixin.qq.com/s/TSNQKkRpN1qbKsT7BvzqIw"],
  "publishedAfter": "30 days",
  "maxArticlesPerAccount": 200
}

Daily monitoring of several accounts (save as a Task and schedule it):

{
  "accounts": ["刀法研究所", "新华社", "https://mp.weixin.qq.com/mp/appmsgalbum?__biz=MzU4Mjg3MDAyMQ==&action=getalbum&album_id=1350949855442862082"],
  "maxArticlesPerAccount": 30,
  "onlyNew": true,
  "stateStoreName": "my-daily-wechat-watch"
}

A whole album as a list of titles, dates and permanent links:

{
  "accounts": ["https://mp.weixin.qq.com/mp/appmsgalbum?__biz=MzIzMDY1MTUxMQ==&action=getalbum&album_id=3346853368002052096"],
  "includeContent": false,
  "sogouPages": 0,
  "maxArticlesPerAccount": 500
}

Output example

A real record from a run on 2026-11-02 (contentText shortened):

{
  "articleKey": "MzU4Mjg3MDAyMQ==_2247638418_1",
  "articleUrl": "https://mp.weixin.qq.com/s?__biz=MzU4Mjg3MDAyMQ==&mid=2247638418&idx=1&sn=...&chksm=...",
  "articleUrlIsPermanent": true,
  "title": "我们最近买的 4 个「新玩意」",
  "digest": "...",
  "author": "少数派",
  "accountName": "少数派",
  "accountId": "gh_...",
  "accountBiz": "MzU4Mjg3MDAyMQ==",
  "messageId": "2247638418",
  "indexInMessage": 1,
  "publishedAt": "2026-11-01T...Z",
  "publishLocation": "广东",
  "coverImageUrl": "https://mmbiz.qpic.cn/...",
  "contentText": "...",
  "contentHtml": null,
  "imageUrls": ["https://mmbiz.qpic.cn/..."],
  "wordCount": 3120,
  "albumId": "1350949855442862082",
  "albumTitle": "少数派的「新玩意」",
  "source": "album",
  "inputAccount": "https://mp.weixin.qq.com/mp/appmsgalbum?__biz=MzU4Mjg3MDAyMQ==&action=getalbum&album_id=1350949855442862082",
  "contentStatus": "ok",
  "collectedAt": "2026-11-02T08:00:00.000Z"
}

Each run also writes a SUMMARY record to its key-value store: per-account counts (albums found, Sogou matches, articles saved) and request statistics.

How the account history is found

WeChat has no public "all articles of an account" page outside the WeChat app. This Actor combines the two public sources that exist:

1. Albums (合集). Many accounts file their articles into albums. Album pages are public and list every article with a permanent link, newest first. The Actor finds an account's albums from its article pages and reads each album to the depth you need.
2. Sogou's WeChat index. Sogou indexes a sample of each account's articles. The Actor searches the account name and keeps only the articles whose account matches exactly.

Both sources are merged and sorted newest first before your limit is applied. Duplicates are removed by article identity (articleKey), not by title: two different articles are both kept even when their titles and publish times are identical. With full text on, a Sogou row's page is read first and its articleKey is compared; a title match alone never removes it. With full text off, Sogou rows whose page was already read to identify the account keep their articleKey and are compared the same way. Only a Sogou row whose id is unknown (full text off, or its page blocked or unreadable) falls back to matching by account, title and exact publish time; see Tips and limits.

Tips and limits

• Coverage depends on the account. Accounts that use albums give a complete, ordered history for those albums. For accounts without albums, you get the Sogou-indexed sample: often the account's best-known articles rather than the newest, and roughly up to 100 per name. Check source on each row.
• Name the account by a link when you can. An article or album link identifies the account exactly. Display names must match the account's name exactly; names shared with many other accounts' article titles (for example a big newspaper) may find fewer matches.
• __biz or gh_ ids alone cannot be looked up without the WeChat app. Paste a link to one of the account's articles instead.
• Temporary links. Articles found only through Sogou get a signed link that expires after some hours (articleUrlIsPermanent: false). Their full text is read in the same run, and articleKey stays stable.
• Some Sogou-only rows have no articleKey. A Sogou signed link does not contain the article id, so the id is known only after the page is read. When the page is not read (includeContent off, except the few newest rows read to identify the account) or cannot be read (contentStatus: "blocked" or "unavailable"), articleKey, messageId and indexInMessage are null. Such rows are deduplicated, and remembered by onlyNew, by account, title and exact publish time. This fallback is approximate: two different articles with the same title posted at the same second would be treated as one. It never overrides a known articleKey. Album rows always have articleKey. Keeping includeContent on gives a key to every Sogou row whose page opens.
• Pasted article links work in the short form (https://mp.weixin.qq.com/s/<code>) and in the long form with chksm. Long links without chksm are usually stopped by WeChat's verification page.
• Blocked pages. WeChat sometimes shows a verification page. The Actor retries from a fresh IP; if all attempts fail, the row is kept with contentStatus: "blocked" and charged at the metadata price.
• Deleted or restricted articles come back as contentStatus: "unavailable".
• Image-only posts have an empty contentText and their pictures in imageUrls.

Integrations

• Daily archive to Google Sheets or a database. Schedule a Task with onlyNew and send new rows to Google Sheets, Airtable or your warehouse through Apify integrations or webhooks.
• Slack or email alerts. Use the run-succeeded webhook to post new titles and links for a watch list of accounts.
• Search and RAG. Load contentText, title, accountName and publishedAt into a vector store; use articleKey as the document id so reruns update instead of duplicating.
• Translation. Pipe contentText into a translation step for non-Chinese-speaking teams.
• Make, Zapier, n8n. Start the Actor and read the dataset from any of these platforms.

Use the API

cURL

curl -X POST "https://api.apify.com/v2/acts/automation-lab~wechat-official-account-articles-scraper/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"accounts":["刀法研究所"],"maxArticlesPerAccount":20}'

JavaScript

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: process.env.APIFY_TOKEN });
const run = await client.actor('automation-lab/wechat-official-account-articles-scraper').call({
    accounts: ['刀法研究所'],
    maxArticlesPerAccount: 20,
    onlyNew: true,
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items.map((a) => `${a.publishedAt} ${a.accountName}: ${a.title}`));

Python

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
run = client.actor("automation-lab/wechat-official-account-articles-scraper").call(
    run_input={"accounts": ["刀法研究所"], "maxArticlesPerAccount": 20}
)
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item["publishedAt"], item["accountName"], item["title"])

Use with AI agents (MCP)

Add this Actor to Claude Code:

claude mcp add --transport http apify \
  "https://mcp.apify.com?tools=automation-lab/wechat-official-account-articles-scraper"

Claude Desktop, Cursor or VS Code (JSON configuration):

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com?tools=automation-lab/wechat-official-account-articles-scraper"
    }
  }
}

Selecting the Actor also adds the run, dataset and key-value-store helper tools. Start the Actor once and keep the returned run ID. If the call returns before the run finishes, check that same run with get-actor-run using a short backoff (for example 2, 4, 8 seconds, capped at 10) up to a deadline you choose, such as 120 seconds. Do not start a new run just to poll.

When the run has succeeded, read results with get-dataset-items:

• Request shape. Pass the dataset ID, an explicit limit and offset, and only the fields you need, for example title,accountName,publishedAt,articleUrl,articleKey. Leave out contentText and contentHtml unless the task needs the article bodies.
• Total budget. Set both a row budget and a byte budget before you start, for example at most 100 rows and 60,000 bytes of serialized UTF-8 JSON in the model context. Page 20 rows at a time and stop at whichever budget is reached first.
• Oversized pages. A single full-text article can be 10–30 KB. If one page would go over the remaining byte budget, request it again with fewer rows (down to 1) or without contentText. Do not truncate a row silently.
• Continuation. Keep the next offset and the dataset's total item count. Tell the user how many rows you read out of how many, and that the rest can be read by continuing from that offset.
• Large results. For more rows than the budget allows, give the user a dataset export link (JSON, CSV or Excel) instead of pulling everything into the conversation.
• Partial results. If a run fails or is still pending at your deadline, report that with the run ID instead of presenting partial rows as complete. A run that stopped at its spending limit says so in its status message; report the row count as partial.

Example prompts:

• "Get the 10 newest articles from the WeChat account 刀法研究所 and summarize each in one sentence."
• "Which articles did 少数派 publish in the last 7 days? List titles and links."
• "Read this WeChat article link and translate the first paragraph into English."

Is it legal to scrape WeChat Official Account articles?

This Actor reads only public pages: public album listings, Sogou's public search results and public article pages. It does not log in, does not use the WeChat app and does not access private messages, follower data or paid content. Articles are copyrighted by their publishers: use the data for research, monitoring and analysis, respect the publishers' rights when republishing, and follow the laws that apply to you, including China's Personal Information Protection Law when articles mention people. This Actor is not affiliated with Tencent, WeChat or Sogou.

Data handling and diagnostics

Failed runs send sanitized diagnostic input, exceptions and Actor/build/run IDs to our private GlitchTip service so we can repair problems. Secret fields and URL queries are removed; reports are retained for 30 days. Raw WeChat and Sogou pages and Sogou cookies are not product output and are kept in memory only for the run. Checkpoints (when onlyNew is on) are stored in a key-value store in your own Apify account. Apify stores your dataset, input and logs under your account's retention settings; delete them in Console when no longer needed. The Actor does not use AI at runtime and does not send scraped data to any AI provider.

FAQ

Can I get every article an account ever published? Only what is public. Album articles are complete for each album. Articles outside albums come from Sogou's index, which holds a sample. The full private history list exists only inside the WeChat app.

Why did my account name return nothing? The name must match the account's display name exactly, and Sogou must have indexed some of its articles. Paste a link to any article of the account instead: that works even when the name search does not.

Why are some links marked as not permanent? Articles found only through Sogou come with a signed link that expires. The full text is read in the same run, so you keep the content; articleKey identifies the article permanently. With full text turned off, or when a page is blocked or unreadable, a Sogou-only row has no articleKey; see Tips and limits.

How does "Only new articles" work with a limit? The first run delivers up to your limit, newest first. Later runs skip delivered articles, deliver new posts first and then continue further back in each album, so a series of runs walks the whole album history. An album is marked as finished only once every article in it has been delivered, so a limit that stops partway through the last listing page loses nothing. An article delivered once is never delivered or charged again for that account, even if a later run finds it through a different source.

Single article links (the article links field) are not checkpointed and are read on every run.

Some rows say contentStatus: "blocked". What happened? WeChat showed its verification page on every retry for that article. The row is kept with title, date and link, charged at the metadata price. Running again later usually reads it.

If such an article was found only through Sogou, its articleKey is null, because the article id is read from the page.

Does it return read counts, likes or comments? No. Those numbers are shown only to logged-in WeChat app users.

Related scrapers

• Sogou WeChat Article Search Scraper: keyword search across all WeChat Official Accounts through Sogou. Use it to discover accounts and topics, then follow the accounts here.