Scientific-Papers

Scientific-Papers is an MCP server that aggregates access to six major scholarly sources (arXiv, OpenAlex, PMC, Europe PMC, bioRxiv/medRxiv, CORE) to provide LLMs with up-to-date papers, metadata, and text content. It offers a suite of tools via a CLI and programmatic commands to fetch latest papers, search across sources, retrieve full texts, perform citation analyses, and resolve DOIs. You can connect an MCP client (e.g., Claude Desktop) using the provided MCP configuration example, or run the CLI locally if you have the server running. The server emphasizes smart rate limiting, cross-source category discovery, and robust text extraction, enabling powerful research workflows for AI assistants and researchers.

To use the MCP server, start it through the recommended npx-based package, then configure your MCP client with the server name scientific-papers. The available tools include listing categories, fetching latest papers, retrieving top-cited papers, searching papers across sources, and fetching specific paper content. The CLI interface is exposed via node dist/cli.js with commands like list-categories, fetch-latest, fetch-top-cited, search-papers, and fetch-content, each supporting source-specific parameters for targeted queries.

Prerequisites:

• Node.js and npm installed on your system
• Basic familiarity with running npm-based projects

Installation steps:

1. Install dependencies and build the server:

   npm install
   npm run build
   

2. Run the MCP server (via the recommended npx approach):

   # Using npx (as per the MCP client instructions)
   npx -y @futurelab-studio/latest-science-mcp@latest
   

3. Optional: If you need to configure a client, use the provided MCP client configuration snippet in your client settings, replacing the server name if desired. The server exposes CLI tools as described in the Usage section.

Prerequisites recap:

• Node.js (recommended latest LTS) and npm
• Internet access to fetch the MCP package via npx or locally built distribution

Tips and notes:

• The MCP server supports both MCP protocol access and a CLI interface for local testing (node dist/cli.js ...).
• Use the npx option for AI tools like Claude Desktop to simplify integration without global installs.
• The server integrates six data sources; ensure you comply with rate limits and respect source-specific usage policies.
• When fetching content, the text body may require additional extraction steps; use fetch-content with appropriate IDs to retrieve full text.
• If you customize the MCP client, refer to the configuration examples in the README to ensure proper command and argument shapes for your environment.
• Environment variables can be added under env in mcp_config if you need to supply API keys or rate-limit settings per source.