Provides CNKI paper search, detail retrieval, and best-match title finding for AI agents via a fast MCP interface.
Configuration
View docs{
"mcpServers": {
"h-lu-cnki-mcp": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/h-lu/cnki-mcp",
"cnki-mcp"
]
}
}
}You run an MCP server that lets Cursor and Claude Desktop directly search CNKI for papers, fetch detailed records, and quickly find the best-matching titles. This enables AI agents to access CNKI content efficiently without manual data gathering.
After you start the CNKI MCP Server, you connect your MCP client (Cursor or Claude Desktop) to the server using the provided local process. You can then perform practical actions such as searching CNKI papers by topic, keywords, or author; retrieving full details for a selected paper; and finding the best-matching paper title for a given query. Use these patterns to power natural-language prompts to your AI agents.
Prerequisites you need before running the server are minimal: ensure you have the MCP runtime tool installed and the Chrome browser available since some automation dependencies rely on Selenium.
Start the CNKI MCP Server using the MCP client tool with the provided configuration. The runtime command is shown below and is copied exactly from the setup guidance:
{
"mcpServers": {
"cnki": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/h-lu/cnki-mcp",
"cnki-mcp"
]
}
}
}Configure your MCP client (Cursor or Claude Desktop) to load the CNKI MCP Server. Use the same runtime command and arguments shown above, but place them in your client’s MCP connection settings.
After configuration, restart the client. Your CNKI server connection should appear as an available MCP endpoint named cnki.
You can add the following configuration blocks in your Cursor or Claude Desktop setup to register the CNKI MCP Server. Use the exact JSON snippets as shown.
{
"mcpServers": {
"cnki": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/h-lu/cnki-mcp",
"cnki-mcp"
]
}
}
}Search CNKI papers by topic, keywords, author, and other supported types to return matching results.
Fetch detailed information for a selected CNKI paper, including title, abstract, authors, affiliation, and DOI.
Quickly identify the most similar CNKI paper title to a given query to streamline discovery.