Chinese Crypto News Radar

One dataset of Chinese-language crypto news from seven sites, with the original text kept as published.

Source key	Site	Script
blocktempo	BlockTempo 動區動趨	Traditional
blockcast	Blockcast 區塊客	Traditional
zombit	Zombit	Traditional
cryptocity	Crypto City 加密城市	Traditional
panews	PANews	Simplified
chaincatcher	ChainCatcher 链捕手	Simplified
techflow	TechFlow 深潮	Simplified

Each row says which site it came from, when it was published, which script it is in, and which ticker symbols it mentions.

What you get that a plain feed reader does not give you

• Ticker mentions checked against a real instrument list. Every row lists the ticker-like symbols found in its title and summary. verifiedTickers keeps only the ones listed on OKX in the same run, so acronyms such as CME, NASDAQ or CPU drop out.
• Crypto and stocks kept apart. OKX also lists tokenised stocks and commodities. tickerAssetClass gives OKX's own asset class for each listed symbol, and cryptoTickers keeps only the ones OKX classes as crypto. A story about Micron earnings gives MU with class stock, not a coin.
• Cross-source mentions. crossSourceTickers and crossSourceVerifiedTickers name the symbols that two or more of your selected sites mention in this run.
• Filters that work in Chinese and English. Keep only some sites, only items that contain a keyword, or only items that mention a ticker you watch.
• A run summary that says what it can and cannot tell you. For each site: whether it answered, how many items its feed carried, how many were dropped for a missing date or link, and how many survived your filters.

Input

Field	What it does
sources	Any of the seven source keys above. Default is all seven. Pick only the Traditional or only the Simplified ones for a single script.
maxAgeHours	Only items published within this many hours. 0 means no limit. Default 72.
maxItemsPerSource	Cap per site after the age filter. Default 50.
keywords	Keep only items whose title or summary contains one of these words.
tickers	Keep only items that mention one of these symbols, for example SOL, PEPE.
allowPartial	On by default. See below.

Output

One dataset row per item, newest first:

Field	Meaning
source	One of the seven source keys
publishedAt	UTC time the site gave, ISO 8601
ageHours	Hours between publication and this run
title, summary	Original Chinese text. summary is empty when the feed gives none, and is cut at 300 characters
url	Link to the item on the site
categoryLabel	The site's own tags when its feed has them
script	traditional or simplified: the script that site publishes in
tickers	Ticker candidates found in the title and summary
tickerSourceCount	For each ticker, how many distinct sites mention it in this run
crossSourceTickers	The tickers of this row that two or more sites mention
verifiedTickers	The candidates listed on OKX (spot or perpetual), checked against OKX's public instrument lists in the same run
tickerVenues	Which OKX list each verified ticker is on: okx:spot, okx:swap or both
tickerAssetClass	OKX's own asset class for each verified ticker: crypto, stock, commodity, forex, bond, unknown, or mixed when the two lists disagree
cryptoTickers	The verified tickers whose class is crypto
crossSourceVerifiedTickers	The verified tickers of this row that two or more sites mention

The run summary is stored under the key SUMMARY in the key-value store.

Things to know before you rely on it

• Tickers are candidates, not a verified list. They are found by pattern in the text. An all-caps word can be a coin, an acronym or a brand. Common acronyms and stablecoins are filtered out, but the filter is not complete. Glued pair symbols such as BTCUSDT are reduced to the base symbol (BTC). Tokenised-stock symbols that begin with a lowercase x are not detected; the title still carries them.
• Listed does not mean the article is about that asset. An acronym can collide with a listed symbol. For example, a story about the Philippine central bank (BSP) matches a symbol OKX lists. Check the title before acting on one.
• Only OKX's lists are used for verification. A coin that trades elsewhere but not on OKX is not verified. If the lists cannot be fetched, all verified fields are null (unknown), not empty, and SUMMARY.listingCheck says which list failed.
• crossSourceTickers is not a signal that something is true or tradable. It only says several of your sites mention the same symbol. Large coins such as BTC and ETH appear nearly everywhere, so look at the less common names.
• These are RSS feeds and carry only their latest items. Some sites keep about 5 or 10 items in the feed, others several hundred. Raising maxAgeHours does not reach further back than the feed itself.
• Each site publishes in one script. Nothing is translated or converted between Traditional and Simplified.
• Some sites could not be included. ABMedia 鏈新聞 refuses requests from the cloud region this Actor runs in, so it is not offered.
• Exchange notices are not included. From the cloud region this Actor runs in, Bybit refuses the requests and OKX serves its US site's notices instead of the Chinese ones, so both were left out rather than worked around.
• Binance is not included. Its robots.txt disallows the announcement API for generic crawlers.

If a source fails

By default the run still returns the sites that answered when one fails. The status message then contains DEGRADED, failed: and the failed site's key, SUMMARY.degraded is true, and SUMMARY.failedSources and SUMMARY.sources give the error. If you would rather have the whole run fail when any site fails, turn allowPartial off. If every selected site fails, the run fails either way.

A site whose newest item is more than 14 days old is treated as broken rather than as a quiet day. So is a feed with no items, or one where no item has a usable date and link.

Data and fair use

The Actor reads public RSS feeds and OKX's public instrument lists, one request each per run. It does not log in, does not use cookies, does not use proxies and does not work around any blocking. It sends a descriptive User-Agent. If a site starts refusing requests, that site is reported as failed; the Actor does not retry aggressively.

News headlines and summaries belong to their publishers. Use them according to the terms of the site you read.