概要
What is Mcp Server To Control Openrefine?
A Model Context Protocol (MCP) server that provides a typed, discoverable interface to OpenRefine’s HTTP API. It lets any MCP-capable client, such as Claude Desktop, orchestrate data-cleaning pipelines safely and reproducibly.
How to use Mcp Server To Control Openrefine?
Install Python 3.13+, uv, and ensure an OpenRefine instance is running (default: http://localhost:3333). Clone the repository, run uv sync, then configure your MCP client (e.g., Claude Desktop) by editing its configuration file to point to the server script and set the OPENREFINE_URL environment variable.
Key features of Mcp Server To Control Openrefine
- Creates projects from dataset URLs
- Applies editing operations to existing projects
- Exports project data as CSV
- Deletes projects
- Provides project model information via a resource URI
Use cases of Mcp Server To Control Openrefine
- Automating data cleaning steps in a reproducible pipeline
- Integrating OpenRefine with AI assistants for conversational data wrangling
- Applying batch operations across multiple OpenRefine projects
FAQ from Mcp Server To Control Openrefine
What is the purpose of Mcp Server To Control Openrefine?
It exposes OpenRefine's HTTP API as MCP tools and resources, enabling AI agents and other MCP clients to interact with OpenRefine programmatically.
What are the prerequisites for running the server?
Python 3.13 or higher, the uv package manager, and a running OpenRefine instance (default URL http://localhost:3333).
How do I set it up with Claude Desktop?
Create or edit the Claude Desktop configuration file to add an MCP server entry that runs uv run openrefine_mcp/openrefine_server.py from the project directory, setting the OPENREFINE_URL environment variable.
Which OpenRefine API endpoints are implemented?
The server implements create-project-from-upload, get-models, apply-operations, export-rows, and delete-project. Endpoints like set-project-metadata, set-project-tags, get-all-project-metadata, preview-expression, and get-processes are not implemented.
Are all OpenRefine features available through this server?
No. The server implements a subset of the OpenRefine API. The README explicitly lists several endpoints as not implemented (marked with ❌).
