OpenRefine MCP Server

A Model Context Protocol (MCP) server that provides a typed, discoverable interface to OpenRefine's HTTP API. This allows any MCP-capable client (like Claude Desktop) to orchestrate data-cleaning pipelines safely and reproducibly.

Installation

Prerequisites

• Python 3.13 or higher
• uv package manager
• OpenRefine instance running (default: http://localhost:3333)

Install the Package

# Clone the repository
git clone <repository-url>
cd openrefine_mcp

# Install dependencies using uv
uv sync

Setup Claude Desktop

1. Create or edit your Claude Desktop configuration file:

   # On macOS/Linux
   ~/.config/claude_desktop_config.json
   
   # On Windows
   %APPDATA%\claude_desktop_config.json
   

2. Add the OpenRefine MCP server to your configuration:

   {
     "mcpServers": {
       "openrefine": {
         "command": "uv",
         "args": [
           "--directory",
           "path/to/your/openrefine_mcp",
           "run",
           "openrefine_mcp/openrefine_server.py"
         ],
         "env": {
           "OPENREFINE_URL": "http://localhost:3333"
         }
       }
     }
   }
   

3. Restart Claude Desktop to load the new MCP server.

Features

This MCP server implements the following OpenRefine API endpoints:

Available Tools

• create_project(dataset_url: str, name: str | None = None) → Creates a new OpenRefine project from a dataset URL
• apply_operations(project_id: int, operations: str) → Applies operations to an OpenRefine project
• export_csv(project_id: int) → Exports CSV data from an OpenRefine project
• delete_project(project_id: int) → Deletes an OpenRefine project

Available Resources

• openrefine://project/{project_id}/models → Returns structural information about the project including column definitions, record model configuration, available scripting languages, and overlay models

Development

Running Tests

make test

Running the MCP Inspector server

make inspector

License

This project is licensed under the MIT License - see the LICENSE file for details.