OpenAI MCP Server logo

OpenAI MCP Server

by thadius83

Query OpenAI models directly from Claude using MCP protocol. This fork adds support for o3-mini and gpt-4o-mini models with improved message handling.

View on GitHub

Last updated: N/A

What is OpenAI MCP Server?

The OpenAI MCP Server is a tool that allows Claude to query OpenAI models directly using the MCP (Message Communication Protocol). It acts as a bridge, enabling Claude to leverage OpenAI's language models for various tasks.

How to use OpenAI MCP Server?

To use the server, you need to install it either via Smithery or manually by cloning the repository and configuring Claude Desktop with the correct server settings, including the OpenAI API key and PYTHONPATH. Once configured, you can use the ask-openai tool within Claude using the <use_mcp_tool> tag, specifying the server name, tool name, and the query along with the desired model (o3-mini or gpt-4o-mini).

Key features of OpenAI MCP Server

  • Direct integration with OpenAI's API

  • Support for o3-mini and gpt-4o-mini models

  • Configurable message formatting

  • Simple interface through MCP protocol

Use cases of OpenAI MCP Server

  • Answering questions using OpenAI models from within Claude

  • Generating text using specific OpenAI models

  • Integrating OpenAI's language capabilities into Claude workflows

  • Comparing responses from different OpenAI models (o3-mini vs gpt-4o-mini)

FAQ from OpenAI MCP Server

How do I install the OpenAI MCP Server?

You can install it either via Smithery or manually by cloning the repository and following the configuration steps.

What models are supported?

The server supports o3-mini and gpt-4o-mini models.

How do I configure Claude Desktop to use this server?

You need to add the server configuration to your claude_desktop_config.json file, including the command, arguments (with your OpenAI API key), and PYTHONPATH.

What is the ask-openai tool?

It's the tool provided by the server that allows you to query OpenAI models directly from Claude.

What do I do if the server is not found?

Verify the PYTHONPATH in your configuration, ensure Python and pip are properly installed, and try running the server directly to check for errors.