WeChat Scraper
Social MediaCollect public data from WeChat official accounts and their articles, finding accounts by name and reading profile and publication details. Agents monitor Chinese-market publishing without an account on the platform.
What This Integration Enables
Every other integration in this part of the catalog publishes something. This one reads. WeChat official accounts are where a large part of Chinese-market product marketing, competitor announcements and industry commentary actually happens, and most Western monitoring stacks cannot see any of it. The connector closes that gap without needing an account on the platform: resolve a company name to an account, read what it published today or across its entire history, pull an article's readable text or its raw HTML, and record how many people read, liked, shared and saved it.
Two things make this intelligence rather than scraping. The first is provenance. Get Official Account Registration Details returns the registering company, the verification status, the authorized third-party operators and the account's history of name changes, which is how you find out that the account you have been benchmarking for six months is run by an agency, or was called something else in March. The second is the comment threads, where the customers of the company you are watching say what they think in public. An agent can archive all of it into S3 or index it into Pinecone for retrieval, but the raw collection is the easy half. The judgment about what a number means, in a market where the numbers have known ceilings, is the half that stops.
Without FlowRunner
With FlowRunner
Use Case Scenarios
Daily competitive archive
The agent holds a list of tracked biz identifiers, resolved once with Search Official Accounts. Every morning it calls Get Daily Article List per account, which returns live rather than cached results and reports each article's delivery status, including the reason when an article was removed or rejected by the platform. That last field is often the most interesting thing in the run. New articles get their text pulled with Get Article Content and their raw markup with Get Article HTML where a faithful archive matters, and the whole set lands in object storage with the publish timestamp and the account it came from. A competitor quietly deleting a post is a signal, and it only registers if you were reading every day.
Campaign benchmarking against the accounts that matter
For a named set of campaign articles, the agent records Get Article Engagement daily to build a read, like and recommendation series per article. For the small set of articles where sharing behavior actually matters, such as a competitor's launch post being benchmarked directly, it uses Get Article Engagement Pro instead, which adds share, save and comment totals at a higher cost per call. The agent checks Get API Balance before a large run and reports the day's spend against yesterday's closing balance, because every call is billed and Get Daily Article List charges even when the account published nothing. The result is a chart the market team can read rather than a screenshot somebody took once.
Sentiment mined from comment threads
The agent takes a set of articles, converts short links to their long form with Convert Article Link, and walks Get Article Comments page by page using the returned buffer value. For threads with substantive discussion it follows Get Article Comment Replies to reconstruct the full conversation rather than only the first level. Comments run through translation with DeepL and get grouped into themes, then land in Notion alongside the article they came from. Where a theme names a specific product problem, the agent raises it in Slack with the original text and the translation side by side, so the person reading it can judge the translation rather than trusting it.
Human-in-Loop Highlight
This connector never writes anything, so the gate is not on the collection. It is one system later, at the moment scraped numbers become a claim somebody acts on. Two specific failures make that gate necessary rather than decorative. First, WeChat caps article read counts at 100001 and the connector reports the cap as-is, so any ranking built on read counts silently ties at the ceiling and "our competitor's post outperformed ours" can mean nothing at all. Second, identity drifts: Search Official Accounts matches fuzzily and returns every account whose name contains the keyword, and Get Official Account Registration Details carries a name-change history and the list of third parties authorized to operate the account. The account you have tracked since March can now belong to somebody else. So before a competitive brief goes to a partnership or ad-buy decision, the agent stops and posts what it knows: "Account tracked as Northwind since March, biz MzI4. Registration now shows a different registering company and one recorded name change on 12 July. Three of this week's five articles report reads at 100001, which is WeChat's ceiling and not a measurement. Confirm this is still the same competitor before I file?" The analyst answers in one message. The cost of skipping that message is a decision made on a number that was never real, about a company that was never the one you meant.
Agent Capabilities
14 actionsOfficial Accounts
3- Search Official Accounts Searches official accounts by keyword and returns each match with its display name, biz and original identifiers, WeChat ID, owning company, category, follower estimate, average headline read and like counts, avatar, QR code, weekly publishing volume and last publish time. Used to resolve an account name into the identifier the article operations need. Fuzzy matching returns every account whose name contains the keyword; exact matching returns only a literal match.
- Get Official Account Profile Returns the public profile of one account: display name, biz identifier, WeChat ID, account type such as subscription or service, avatar and profile description. Accepts a name, an original ID or a WeChat ID, which makes it the quickest way to confirm an identifier points at the account you expect before a bulk run.
- Get Official Account Registration Details Returns the registration record behind an account: the registering company or organization, registration date, verification status and customer type, registered country and province, the third parties authorized to operate it, and any history of name changes. This is the provenance check that keeps a monitoring program honest.
Article Listings
3- Get Daily Article List Returns everything an account published today, live rather than cached, covering up to 20 publishing batches of up to 8 articles. Every entry carries the link, title, summary, cover image, publish timestamp, position in its batch, original or reposted flag, content type and delivery status including the reason an article was removed or rejected. The account is charged even when nothing was published.
- Get Historical Article List Returns an account's publishing history one page at a time, live rather than cached, with five batches per page and up to eight articles per batch. The response reports total pages, and walking the whole history means repeating the call with the offset returned by the previous page. Since December 2024 this endpoint returns short links.
- Convert Article Link Converts between the short and long forms of a WeChat article link and returns both. Historical listings return short links while most engagement and comment operations need the long form, so this is the bridge between them.
Article Content
2- Get Article Content Returns an article's body as readable text with its metadata: title, long link, hash ID, publishing account name and identifiers, publish and scheduled times, position in its batch and the images it contains. Choose plain text with image tags for clean text that keeps image placeholders, or rich text HTML for markup that drops into an editor.
- Get Article HTML Returns the raw HTML body exactly as published, alongside title, canonical URL, cover image, publishing account details and publish time. Used when the original markup matters, such as archiving an article faithfully or running your own parser over it.
Engagement
2- Get Article Engagement Returns the three headline figures for an article: read count, like count and "wow" recommendations. Read counts are capped at 100001 by WeChat and reported as-is. This is the cheaper engagement call and the right one for tracking a large set of articles over time.
- Get Article Engagement Pro Returns the full engagement picture: reads, likes, "wow" recommendations, shares, saves and total comments. Costs more per call, so it is reserved for the articles whose sharing and saving behavior actually matters.
Comments
2- Get Article Comments Returns the top level comments on an article, each with text, author nickname and avatar, like count, creation time, pinned state, content ID and any replies the account has already posted. Requires the long form of the link and pages by passing the buffer value from the previous call.
- Get Article Comment Replies Returns the replies under one top level comment, each with text, author avatar, like status, creation timestamp and the commenter's country and province as WeChat reports it. Requires the long link and the parent comment's content ID.
Search
1- Search Articles Searches the indexed archive of WeChat articles and returns matches with title, long and short links, publishing account, publish time, engagement figures and a body excerpt, plus total hits and pages. Three keyword fields combine into precise queries: required terms, any-of terms and excluded terms. Title search reaches back up to 720 days, body search is limited to 30 days, and each page returns 20 results.
Account
1- Get API Balance Returns the remaining credit on the API account with yesterday's closing balance and the time of the reading. Every scraping call is billed per request and fails once the credit runs out, so an agent checks this before a large batch and watches the daily difference to track spend.
Frequently Asked Questions
What can FlowRunner do with WeChat Scraper?
FlowRunner agents can run Search Official Accounts, Get Official Account Profile, and Get Official Account Registration Details in WeChat Scraper, plus 11 more actions.
Does connecting WeChat Scraper to FlowRunner require OAuth?
No. WeChat Scraper connects to FlowRunner with an API key, no OAuth flow required.
Can WeChat Scraper trigger a FlowRunner workflow automatically?
WeChat Scraper doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.
Start building with WeChat Scraper
$100 in credits. No card required. Connect in minutes.