China News Feed with Deletion Watch

Who said what, and what went missing.

Chinese news from many outlets in one dataset, each row labeled by the type of outlet that published it, plus the public hot lists of Weibo 微博, Baidu 百度, Zhihu 知乎, Toutiao 今日头条 and Bilibili 哔哩哔哩. When a topic drops off one of those five lists while it is still hot, we recheck it, flag it and show the evidence for each flag. Douyin 抖音 and Kuaishou 快手 are in the default hot lists but come in as hot-list rows with no deletion check. Kuaishou is paused by default.

What you get

News rows labeled by outlet type	Every news row has a source_type: State media, Party-gov, Commercial platform, Overseas independent, Dissident or Censorship tracker. Aggregator reposts are credited to the original outlet. The label describes the outlet. It says nothing about whether a story is true.
Stories grouped across outlets	Rows about the same story share a topic_id, so you can put state media and overseas coverage of one story side by side.
Hot-list rows	Weibo 微博 hot search, Baidu 百度 hot search, Zhihu 知乎 hot list, Toutiao 今日头条 hot board and Bilibili 哔哩哔哩 hot search, with rank, heat, peak_rank and rank_trail for each topic. Douyin 抖音 and Kuaishou 快手 are in the default hot lists but come in as hot-list rows with no deletion check. Kuaishou is paused by default.
Deletion flags with evidence	Topics that leave those five lists while hot are flagged Likely censored by default; Censored only with a dead link, removal notice or close tracker match. Topics that come back are marked Returned. We show the evidence for each flag. Weaker signals are filtered out before they reach your dataset.
Chinese feed and English feed views	The same rows in two dataset views. The English feed shows headline_en, excerpt_en and summary_en, which are machine translations or generated text. English is added only where a translation is available, and right now most rows are not translated; those rows keep the Chinese and show translation_status: translation_pending.
Price	$8 per 1,000 results ($0.008 per row). No start fee. Deletion checks are included.

Quick start

1. Open https://apify.com/china.unofficial/china-news-feed in Apify Console. Pick the hot lists and source types you want, or keep the defaults.
2. Check Max billed rows (maxItems). The Console pre-fills 50. If Max billed rows is left empty, the cap is 200 new rows. A 50-row run costs at most $0.40. Use the default timeout (600 s) for complete, deduplicated results.
3. Run it. When it finishes, open the dataset in the Flagged stories, English feed or Chinese feed view and export it as CSV, JSON or Excel.

Default run settings are a 600-second timeout and 1024 MB of memory. A run can return fewer rows than maxItems, or none, if it hits its time limit or the sources don't have enough new rows after filtering. Held rows arrive on your next run, and you're billed only for rows delivered.

Statuses

Status (status)	Label	What it takes
likely_censored	Likely censored / 【疑似被审查】	The default label. The topic dropped off abruptly while hot and was still gone at the 10-minute recheck.
censored	Censored / 【已被审查】	Still gone after 1 hour, and only with a dead link (the topic URL returns 404/410), a platform removal notice such as Weibo's 「根据相关法律法规」 notice, a Zhihu 4041 error, or a close match to a specific China Digital Times (CDT) item. provisional stays true until the 24-hour recheck.
returned	Returned	The topic was flagged and then came back. We keep it visible rather than quietly deleting the flag.

Deletion watch covers the Weibo 微博, Baidu 百度, Zhihu 知乎, Toutiao 今日头条 and Bilibili 哔哩哔哩 hot lists. Douyin 抖音 and Kuaishou 快手 are in the default hot lists but come in as hot-list rows with no deletion check. Kuaishou is paused by default. Platforms curate their hot lists, so a topic can leave for commercial or editorial reasons. So the rule is: Likely censored by default; Censored only with a dead link, removal notice or close tracker match. We show the evidence for each flag.

English translations and run status

Each English row has a translation status: ok, translation_pending or failed. A pending row shows "Translation pending; original shown" with the Chinese headline in place of the English one.

When some sources are down, the run's status message says: "Some sources, story matching and English translations are temporarily unavailable in this run."

Two fields worth knowing

• fetch: how this row is refreshed, either scheduled (each run, the default) or hourly.
• via: the third-party copy a hot-list row came from, or null.

Pricing

• $8 per 1,000 results, which is $0.008 per row.
• Pay per event: one billed event per dataset row. No start fee.
• Deletion checks are included. Flag rows are billed like any other row and count toward maxItems.
• 50 rows cost $0.40.
• Emit only new rows (emitOnlyNew) is on by default, so rows you already received aren't billed again. A flag comes back as a new row when its status changes, for example from Likely censored to Censored, or to Returned. Run one at a time to avoid repeats.

FAQ

Why "Likely censored" rather than "Censored"? Likely censored is the default label. A flag becomes Censored only with a dead link, a platform removal notice (such as a Weibo notice or a Zhihu 4041 error) or a close match to a specific CDT item.

Can a Censored flag change? Yes. It stays provisional until the 24-hour recheck. If the topic comes back, its status becomes returned.

Does a source label mean a story is true or false? No. source_type describes the outlet, not the story.

Am I charged for deletion flags? Each flag is one dataset row at the same $0.008 price, and it counts toward maxItems. There is no separate deletion fee.

Why did I get fewer rows than Max billed rows? The run hit its time limit, or the sources didn't have enough new rows after filtering. Held rows arrive on your next run, and you pay only for rows delivered.

Can I filter by date, keyword or language? No. The Actor has no date-range, keyword or language input. To get only what's new, schedule it with Emit only new rows left on. You can narrow a run with hotLists, sourceTypes, newsSources and maxHotRank.

Is every row in English? No. English headlines and excerpts are machine translations added only where available. Right now most rows ship in Chinese only, with translation_status: translation_pending. Check names and slang against headline_original.

Which platforms does the deletion watch cover? Weibo 微博, Baidu 百度, Zhihu 知乎, Toutiao 今日头条 and Bilibili 哔哩哔哩 hot lists. Douyin 抖音 and Kuaishou 快手 are in the default hot lists but come in as hot-list rows with no deletion check. Kuaishou is paused by default.

Leave a review

If this Actor is useful in your work, a rating or short review on the Store page helps other people find it. If something is wrong or missing, the Issues tab is the best place to tell us.