core_ac_uk_mcp

chkunkel/core_ac_uk_mcp

3.1

If you are the rightful owner of core_ac_uk_mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

This project provides a minimal MCP server for accessing the CORE.ac.uk API, enabling local programmatic access to academic content.

Minimal MCP Server for Accessing CORE.ac.uk API

This project provides a lightweight Machine Communication Protocol (MCP) server that wraps key functionalities of the CORE.ac.uk API. It enables local programmatic access to academic content, allowing you to search works by keyword or download full metadata for individual research papers using a simple interface.

Ideal for integrating into applications or development environments that support MCP-based extensions.

API documentation: https://api.core.ac.uk/docs/v3

Installation

  1. Clone repository:

    git clone https://github.com/chkunkel/core_ac_uk_mcp
    cd core_ac_uk_mcp
    
  2. Install dependencies with uv:

    uv sync
    

    When installed, run:

    uv run <path_to_directory>/core_mcp_server_local.py
    

    Please add the path to the file, you don't need quotes or the < > brackets around your key.

    Output (if successful):

    INFO     Starting MCP server 'CORE\xa0API\xa0Server' with transport 'stdio'
    

    With an API key installed in the environment, the MCP server is ready.

API Key Setup

You need to obtain an API Key on the website first.

Then create a .env file in the main directory with the following content:

CORE_API_KEY=<your key>

You don't need quotes or the < > brackets around your key.

Server Configuration

Run in console as python core_mcp_server.py

Add the following config to the program of your choice:

{
  "mcpServers": {
    "core_ac_uk_server": {
      "command": "uv",
      "args": [
        "run",
        "<SCRIPT_PATH>/core_mcp_server_local.py"
      ],
      "env": {},
      "working_directory": null,
      "start_on_launch": true
    }
  }
}

Replace <SCRIPT_PATH> by the respective path to the script

Available Functions (so far)

  • search_works_by_keywords
  • download_single_work_by_identifier