Can I retrieve a full WeChat article, and why did my AI only read its summary?
Yes. Request the article URL with wechat_mp.v2.fetch_article_detail_h5. The response puts the plain-text body in data.content.content_text, the HTML body in data.content.content_noencode and the summary in data.content.desc. An assistant that answered from the summary received desc or a tool preview instead of content_text. To ask a question about the article, give the assistant content_text and the original article link.
Fetch the text from one article URL
Pass the complete mp.weixin.qq.com/s/… short link or mp.weixin.qq.com/s?__biz=… long link as url. If you already saved a successful response for the article, read the saved response instead of requesting the article again.
Check whether the response contains the passage you need
- Check that the request succeeded, then verify the returned title, the account display name in
nick_nameand the article link indata.content.link. Keep the byline fromauthor, and keep any other source the author supplied fromsource_url. - Look for
content_text. If the assistant only receiveddescor a tool preview, read the saved result of the same request. With Capabilities or MCP, setfieldson the result selection.wechat_mp.v2.fetch_article_detail_h5usesraw:trueby default to retain metadata, and the body stays atdata.content.content_texteither way. - For a registration deadline, find the sentence that contains the date and read the conditions around the sentence. If
content_textis only an introduction, or the date appears in a poster, video or attachment, open the original article and record which materials you reviewed.
For an account’s publishing history, start with its article list. Use article statistics for read counts and the comment workflow for selected comments.
Once the assistant has content_text, try asking: “Find the deadline and who can apply. Include a short quote and the original link. Flag anything you couldn’t verify in the text.” To save responses and check individual quotations, use the article reading guide. BeatAPI pricing
If the article is part of an automation, read the article in n8n and keep the article URL, text and request ID in the output.

