How do I separate WeChat headline articles from secondary positions in a saved account list?
To tell a headline from a secondary article, read idx from the article list returned by wechat_mp.v2.fetch_account_articles: 1 is the headline article of the publication batch, 2 is the second article, and higher values are later positions. Filter idx=1 to study the first article of each push, or keep every position to examine whole batches.
Why isn’t the first row of a result page necessarily a headline?
A result page can contain several publication batches or begin partway through one batch. Placement is the article’s position within the article’s publication batch, not the article’s row on the current page.
For example, two batches on one page:
Filtering yields two headlines and one secondary article. Keep both batch A records even though the two records share app_msg_id: article identity combines account, app_msg_id and idx.
Collect a headline-only reading list
- Retrieve the verified account’s article list, selecting the article tab and
raw:false. - Read
data.articles[].idxand keep the articles with position 1, with title, original link and publication time. - Save articles with positions above 1 separately for comparison, and mark records without a position as pending review.
Reuse saved list responses; do not request every body only to determine placement, because the list’s idx field already gives the placement. If you start from one article URL, use the article-to-account handoff to find the account, then locate the article in the account’s list.
To hand the task to an assistant connected to BeatAPI, use this instruction:
Organize the latest three list pages of the target account. Keep headline articles with title, link and publication time, save secondary articles separately and flag records with missing positions. Report the actual number of headlines collected.
Read the files that article_positions.py exports
article_positions.py (in the last section) splits saved list pages into four files:
headlines.csv: position 1.secondary.csv: later articles with their exact positions.unknown-position.json: records without a returned position.coverage.json: page sources, observed article distribution and counts.
If a batch shows only position 2, keep the position 2 record and revisit the list when you need the batch’s first article. For engagement research, group headlines and secondary articles first, then compare publication time and subject. Continue with statistics for selected articles. Retrieval prices are on the pricing page.
Optional export from saved lists
Retrieve the lists with data:wechat_mp.v2.fetch_account_articles; article_positions.py then filters the saved responses locally.
article_positions.py accepts 1–20 saved article-tab pages of the same canonical gh_ account, requested with raw:false. Record the actual requests, times and files in the manifest; each later page uses the previous response’s next_offset.
Replace the history-pages.json example below with your own records:
Save the code as article_positions.py, then run:

