WeChat Official Account data APIs: how to choose (2026)

Scraping WeChat yourself means keeping WeChat accounts signed in, getting past captchas and bans, and fixing parsers when pages change; a data API is called with a key. RedFox gives 20 articles with reads for ¥0.04 but a day late, DataXAPI fetches at call time and charges reads per article, and WeChat's own API covers only accounts you run. DataXAPI's endpoints here are from $0.0200 a call, and a failed call is not charged; what others charge and do was checked on 2026-10-29, each with its source below.

Open in ChatGPTOpen in Claude

Five questions and four routes

Before choosing, ask five things: can it read accounts you do not run; does it give reads; is a failed lookup charged; what is the least top-up and in what currency; and is there a contract and an invoice. The table goes through them.

RouteAccounts you do not runReadsFailed lookup charged?Least top-up and currencyContract and invoice
WeChat's official APINo: only with your own or an authorised account's credentialsYes, article read data up to yesterdayFreeFreeNot applicable
Your own scraper (Sogou WeChat search, packet capture)YesBy packet capture, signed in to your own WeChatNot applicable; the cost is accounts and timeNot applicableNot applicable
RedFox (红狐)YesYes, with the article list and article detailsFailed requests (not 200) are not charged¥100 at least, ¥1 = 10 creditsA contract SLA for enterprise users: a ¥3,000 top-up or ¥10,000 spent
DataXAPIYesYes, one call of the stats endpoint per articleFailed calls, timeouts and articles that do not exist are not charged; an article list, search or comments call with no results is an empty result and is charged$10 at least, a US-dollar balance; Alipay paying in yuan, or a card; $0.10 of trial credit on sign-upNo company transfer or invoice for now

Prices and freshness

One common job makes the prices comparable: an account's last 20 articles with their reads. Some APIs put reads in the article list, others charge them article by article, so the unit price and the cost of the job can differ a lot.

RouteArticle listReads of one article20 articles with readsFreshness
WeChat's official APIFree, 1–20 a call, your own account onlyFreeFreeArticle read data: end_date no later than yesterday
Your own scraper (packet capture)Not applicableNot applicableNot applicableIts parameters expire in about 4 hours; the project says it cannot be real time
RedFox (红狐)¥0.04 a call, 20 articles a page with reads, wows and likes; down to ¥0.02 at volumeNot sold apart: it comes with the list¥0.04 (one call)Marked T+1, updated once a day with the day before; the article list goes back to 2026-04-01 at the earliest
DataXAPI$0.0200 a page, up to 20 articles, without reads$0.0200 an article1–2 list pages, plus 20 × $0.0200Fetched at call time; every response carries meta.fetched_at

Where the others are better

RedFox: one call gives 20 articles with reads, wows and likes, at ¥0.04 before discounts. The same 20 articles with reads cost DataXAPI a list plus one call per article, far more. Its data is a day old; if that is fine, it is the cheaper way.

WeChat's official API: for your own account the data is first-hand and free. If your own account is all you need, no third party is needed.

DataXAPI has no company transfer or invoice for now. A company that needs a contract finds one on RedFox's enterprise plan.

How DataXAPI differs

Calls by link, ID or keyword are fetched at call time, and every response carries meta.fetched_at, the moment it was fetched. Above 100,000 reads WeChat shows only “10万+”; the count then comes as 100001, marked a lower bound in approximate.

No WeChat account or cookie of yours is used. The key goes in the Authorization: Bearer header.

Official Accounts, Channels and WeChat Search share one key and one JSON shape with Douyin, Xiaohongshu and the rest. The remote MCP server is https://api.dataxapi.com/mcp; ChatGPT and Claude connectors sign in with OAuth, with no key typed. llms.txt, openapi.json, prices.json and a .md version of each page are public.

Who should pick which

Only your own accounts: WeChat's official API. Reads for many articles at once, data a day old is fine, a tight budget: RedFox.

The figures as of the call, more than Official Accounts (Douyin, Xiaohongshu or platforms outside China), MCP in an AI client, a small top-up in dollars or by Alipay: DataXAPI. A few articles now and then, and time to keep accounts and parameters alive: your own scraper.

How to switch to DataXAPI

An account's article list: wechat_mp.accounts.articles.list, with the original ID starting gh_ as id. Knowing only the name, search its articles with wechat.posts.search and take author.id. For the next page pass page.next_cursor back as cursor, within the hour.

Reads, likes, wows: wechat_mp.articles.metrics, with the long link (__biz, mid, idx, sn) as url; the article list gives long links. metrics.views is reads, metrics.likes is likes, and with view=full, ext.recommends is recommends (once 在看, “wow”).

Text, comments, search: the text from wechat_mp.articles.get, which takes the long link or the short /s/ one; featured comments from wechat_mp.articles.comments.list; WeChat Search from wechat.posts.search, with the terms as query and optionally sort and published_within.

Questions

Can I get the reads of an account I do not run? Not through the official API, which takes only your own or an authorised account's credentials. Third-party APIs can: RedFox and DataXAPI both do.

Is a lookup of a deleted article charged? It depends. RedFox does not charge requests that are not 200; DataXAPI does not charge an article that does not exist.

Which needs my own WeChat account? Only packet capture. The official API needs the Official Account's AppID and AppSecret; the third-party APIs need only their own key.

Endpoints

Each with its price; every page has the parameters, the fields and a recorded response.

Sources

What this page says of other services is from their own pages and documents, read on 2026-10-29. Prices and features change: check the source before you decide.