Chinese literature
Search metadata and abstracts for Chinese journals, theses and conference papers
Introduction
Ai4Scholar searches Chinese journal articles, theses and conference papers, returning metadata and available abstracts. Search is enabled; detail lookup and full-text downloads are currently unavailable.
Current entry points
Choose Chinese literature, after Google Patents, in Literature search. To use it in research conversations, enable it in Settings → Search sources and ask for Chinese literature or research in China.
Developers can open Settings → API docs → Chinese literature API or the interactive reference, then call POST /api/proxy/wanfang/search with an Ai4Scholar API Key.
Data coverage
- Document types: Chinese journal articles, theses and conference papers, with type filters.
- Content: Metadata and abstracts supplied by the source. Missing abstracts are marked as metadata only.
- Full text: An abstract does not mean the full text was retrieved. PDF downloads are not included.
Search suggestions
Use concise Chinese keywords. Queries match titles, abstracts and keywords. Space-separated concepts must all match; double quotes preserve a phrase. For example, 人工智能 教育 combines artificial intelligence and education. Raw PQ expressions, Boolean syntax and vector parameters are not passed through.
Filter by document type or publication years, and sort by relevance, newest year or citation count. If no matches appear, use fewer keywords, synonyms or shorter title fragments.
Result counts and pagination
| Entry point | Display or request count | Maximum fetched per request |
|---|---|---|
| Search page | 10 displayed per page by default, configurable from 1–100 | Up to 100 fetched and cached |
| Ai4Scholar API | limit defaults to 10, configurable from 1–100 | 100 |
| Research conversations | 10 | 20 |
Save a Chinese literature default from 1–100 in Settings → Search sources → Set desired result count. The search filter menu also allows a temporary override. This setting controls the display page size; the search page still fetches up to 100 results per request. Research conversations cap preferences above 20 at 20. Actual results may be fewer than requested.
The API uses limit for the result count and offset for the starting position (0–5000). Keep other parameters unchanged and use the returned nextOffset for the next request; stop when it is null.
Billing
Each successful new search request has a base price of 1 credit, regardless of the result count. Actual pricing depends on account discounts and usage records. Browsing already loaded results within the same search incurs no additional charge. For example, when a request returns 100 results and the display page size is 10, the first 10 pages together incur one search charge. After browsing those results, use Fetch next batch, which shows the cost, to make another request. Each click makes at most one search request. If the last display page is incomplete, new results fill it before advancing so no papers are skipped. See Credits and plans.
The page cache lives only in the current browser tab. Starting a new search, changing query conditions or searching again after a refresh creates a new request.
API clients can retry the same request with X-Ai4S-Idempotency-Key to avoid duplicate charges. Successful results are retained for 24 hours. Pending or expired results return 409; use a new key only for an intentional new search.
How it complements other sources
Use Chinese literature search for reviews and research in China, alongside Semantic Scholar, PubMed and Google Scholar for international literature. Preserve source attribution when saving or citing results. Abstracts support abstract-level evidence, not claims of full-text reading.