Zongheng (纵横中文网) Web Novel Scraper

Scrapes zongheng.com — one of China's largest web-novel platforms — for books, categories, rankings and book metadata.

Data Source

• Search: search.zongheng.com/search/book?keyword=... — clean JSON API
• Category / ranking / detail pages: server-rendered pages with an embedded window.__NUXT__ state blob, decoded by a built-in tokenizer (no browser needed)

No login or API key required.

Modes

Mode	Inputs	Source
search (default)	searchQuery	keyword JSON API, paginated
byCategory	categoryId (8101 玄幻奇幻 … -100 其他)	/categories?cateFineId= curated lists — all 13 pre-rendered sections (click/new-book/hot-sales/recommended/classics/premium/trial-read/banners etc.)
ranking	rankType (monthTicket/popular/newBook/newOrder/click/recommend)	/rank?nav=default — all six lists are pre-rendered in the SSR blob
byBookUrl	bookUrls	/detail/{id} metadata + first/latest chapter info

Output fields

book records: bookId, title, author, authorId, category, subCategory, categoryId, subCategoryId, status (serialized / completed), wordCount, description, keywords, coverUrl (static.zongheng.com, verified 200), latestChapter, updateTime, url. Detail records add recommendCount, clickCount, createTime, latestChapterId, latestDate, firstChapter, firstChapterUrl, latestChapterUrl, authorTotalWords.

Every record carries recordType, sourceUrl and scrapedAt.

Reliability

• Exponential-backoff retries on 429/5xx and intermittent Chinese-CDN timeouts.
• On 403/429 the actor lazily engages the free Apify AUTO datacenter proxy group.
• Invalid book URLs / unparseable pages produce typed recordType: "error" records.

Limitations

• SSR-dependent modes (byCategory, ranking, byBookUrl): zongheng serves the full server-rendered state only to some networks; from some datacenter IPs it may return a client-side shell without the window.__NUXT__ data. Those modes then fail-soft with a typed recordType: "error" record (SSR_UNAVAILABLE / BOOK_UNAVAILABLE) and a status message — retry and they typically resolve (verified working from Apify cloud). The search mode uses a clean JSON API and works everywhere.
• The byCategory mode walks all 13 pre-rendered category sections (click rank, new books, hot sales, recommendations, classics, premium, trial-read, etc.); the serialStatus value inside some category rank lists is zeroed upstream (unreliable), so status is most accurate in search/byBookUrl modes.
• wordCount is omitted when upstream reports 0 (stub value, not a real count).
• Chapter links (firstChapterUrl / latestChapterUrl) point at read.zongheng.com, which is Referer-protected: the links return 403 from a clean client (fresh browser tab, curl, or a programmatic fetch without a Referer) and load only after a visit to the book page on www.zongheng.com. The links are correct — set includeChapters=false if you don't need them.
• Full chapter lists live behind the client-side wwwapi.zongheng.com API which is unreachable from this network; book records include the first and latest chapter names/IDs/links instead. Set includeChapters=false to drop them.
• Category pages expose curated sections per category (not an exhaustive paginated list — the full list is SPA-only).