CNKI MCP Server
Enables searching and retrieving academic papers from CNKI (China National Knowledge Infrastructure), including support for DOI-based searches and retrieval of paper metadata such as titles, abstracts, authors, institutions, and DOI identifiers.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@CNKI MCP Server搜索关于人工智能伦理的最新论文"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
CNKI MCP Server
基于 FastMCP 框架的中国知网 (CNKI) 论文检索 MCP 服务器,使 Cursor、Claude Desktop 等 AI Agent 可以直接搜索和获取 CNKI 论文信息。
功能特性
search_cnki: 搜索 CNKI 论文,支持多种搜索类型(主题、关键词、作者等)
get_paper_detail: 获取论文详情(标题、摘要、作者、机构、DOI 等)
find_best_match: 快速匹配最相似的论文标题
快速使用 (无需手动安装)
本项目的代码已发布到 GitHub,你可以通过 uvx 直接运行,无需先把代码克隆到本地安装。
前置要求:
需安装 uv
系统需安装 Chrome 浏览器(Selenium 依赖)
1. Cursor 配置
在 Cursor 设置中添加 MCP 服务器。编辑 ~/.cursor/mcp.json (或通过 UI 添加):
{
"mcpServers": {
"cnki": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/h-lu/cnki-mcp",
"cnki-mcp"
]
}
}
}2. Claude Desktop 配置
编辑配置文件 (通常位于 ~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": {
"cnki": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/h-lu/cnki-mcp",
"cnki-mcp"
]
}
}
}配置保存后,重启 Cursor 或 Claude Desktop 即可看到 cnki 服务器已连接。
使用示例
在 Cursor / Claude 中直接提问:
帮我搜索关于"大语言模型"的 CNKI 论文
获取这篇论文的详细信息: https://kns.cnki.net/kcms2/article/abstract?v=...
查找这篇论文: "Transformer: Attention Is All You Need" 在知网上的情况
支持的搜索类型
中文 | 英文别名 |
主题 | subject, theme |
关键词 | keyword, keywords |
篇名 | title |
作者 | author |
作者单位 | affiliation, institution |
DOI | doi |
注意事项
建议每次搜索 1-3 页,避免频繁请求
搜索间隔建议 2-3 秒
CNKI 可能有反爬限制,如遇问题请适当降低请求频率
首次运行时会自动下载 ChromeDriver,可能会花费一点时间
License
MIT
Available Tools
3 toolsfind_best_matchA
快速查找与输入标题最匹配的 CNKI 论文。
使用字符匹配算法,适合用于验证论文标题或快速定位特定论文。
Args: query: 论文标题或关键词 ctx: MCP 上下文(自动注入)
Returns: 最匹配论文的标题和 URL
Inputs: query (required). Returns: object
get_paper_detailA
获取 CNKI 论文详情页的完整信息。
Args: url: CNKI 论文详情页 URL(通常从 search_cnki 结果中获取) ctx: MCP 上下文(自动注入)
Returns: 包含论文完整信息的字典:title, authors, institutions, abstract, keywords, source, year, volume, issue, pages, doi, cited_count, download_count, fund
Inputs: url (required). Returns: object
search_cnkiB
搜索 CNKI 论文,返回论文列表。
Args: query: 搜索关键词 ctx: MCP 上下文(自动注入) search_type: 搜索类型,支持: - 中文:主题、关键词、作者、篇名、作者单位、全文、DOI、基金、摘要 - 英文:subject, keyword, author, title, affiliation, fulltext, doi pages: 搜索页数(1-10),每页约20条结果 sort: 排序方式,支持: - 中文:相关度、发表时间、被引、下载、综合 - 英文:relevance, date, cited, download, composite
Returns: 包含论文列表的字典,每篇论文包含:title, url, authors, source, date, cited_count
Inputs: query (required), search_type, pages, sort. Returns: object
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
TDQS
Scored across 3 tools
The three tools have clearly distinct purposes: search_cnki performs broad searches with filtering options, find_best_match uses character matching for precise title verification, and get_paper_detail retrieves detailed metadata from a specific URL. There is no functional overlap between these tools, making selection unambiguous.
All tool names follow a consistent verb_noun pattern (search_cnki, find_best_match, get_paper_detail) with clear, descriptive naming. The conventions are uniform throughout the set, using snake_case and action-oriented verbs appropriate to each tool's function.
With only three tools, the server feels somewhat thin for a comprehensive CNKI (academic paper database) interface. While the tools cover core search and retrieval workflows, typical academic search servers would include more operations like citation analysis, author profiling, or filtering capabilities. The count is borderline minimal but functional.
The tools provide good coverage of essential CNKI operations: searching, precise matching, and detailed retrieval. Minor gaps include lack of update/delete operations (though these may not apply to read-only academic databases) and advanced filtering beyond the search parameters. The surface supports basic research workflows effectively.
Maintenance
Related MCP Connectors
Academic papers, reference verification and research data for AI agents.
Search peer-reviewed papers and research methodology guidance from your AI agent.
arXiv and OpenAlex/Crossref paper search (an open Google Scholar alternative) as agent tools.
Find research papers and arXiv preprints, and get an abstract and free full text from a DOI or id.