Openai Codex MCP Server

@agency-ai-solutions

代码仓库

基本信息

分类

AI 与智能体

许可证

NOASSERTION

传输方式

stdio

作者

agency-ai-solutions

关于 Openai Codex MCP Server

An MCP server to communicated with, use, and wrap the API for the OpenAI Codex CLI tool.

#openai-codex#mcp-server #api-wrapper

Openai Codex MCP Server 配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "openai-codex-mcp": {
      "command": "uv",
      "args": [
        "venv",
        "#",
        "creates",
        "a",
        ".venv/",
        "directory"
      ]
    }
  }
}

Openai Codex MCP Server 工具

工具列表暂未提供

当前收录暂时没有可展示的工具列表，不代表服务没有工具。请查阅项目文档了解支持的能力。

工具是从 README 中自动提取的。维护者可以在 ## Tools 标题下列出工具,即可填充这部分内容。

概览

What is openai-codex-mcp?

openai-codex-mcp is a JSON-RPC server that wraps the OpenAI Codex CLI tool, enabling Claude Code to use OpenAI’s models for code generation, explanation, and debugging within the same interface.

How to use openai-codex-mcp?

Install with uv pip install . after creating a venv, then run codex_server (or uvicorn codex_server:app). Register it in Claude Code via claude mcp add /path/to/openai_codex_mcp.json or manually with base URL http://localhost:8000/ and protocol JSON-RPC 2.0. No authentication needed.

Key features of openai-codex-mcp

Specialized methods: write_code, explain_code, debug_code
Supports multiple OpenAI models (O‑series and GPT)
Simplified parameter naming for easy integration
Runs as a lightweight, standalone Python server
No authentication required for the MCP server itself

Use cases of openai-codex-mcp

Generate code in a specific language for a given task
Obtain detailed explanations of how code works
Find and fix bugs in problematic code
Leverage different OpenAI models (o4‑mini, gpt‑4o, etc.) for specialized tasks

常见问题

What makes this MCP server different from directly using the Codex CLI?

It provides a JSON-RPC interface specifically designed for Claude Code, with dedicated methods for common coding tasks and simplified parameter naming.

What are the runtime requirements?

Python 3.12+, the OpenAI Codex CLI tool (installed via `npm install -g @openai/codex`), and a valid OpenAI API key configured for the Codex CLI.

Where does the code/data processed by the server live?

The README does not specify data storage; all processing occurs locally via the Codex CLI and the MCP server on localhost.

Does the server support authentication?

No – the base URL configuration does not require an authentication token.

What transport protocol does it use?

JSON-RPC 2.0 over HTTP, listening on port 8000 by default.

AI 与智能体分类下的更多 MCP 服务器

Plainpaper

The AI marketing canvas for Claude and MCP agents. Your agent drafts campaigns as cards on a shared board, briefed by your brand rules, approved by you.

litshrine

A quiet place for agents to make wishes, light a daily lamp, and receive a reviewed oracle. Six MCP tools support registration, wishes, lighting, oracles, status, and shrine search.

UnificAlly MCP

unifically-dev

UnificAlly MCP server. One API for 100+ AI video, image, music and speech models. https://mcp.unifically.com/mcp

CrawlForge MCP Server

mysleekdesigns

28-tool MCP server for web scraping, crawling, deep research & autonomous extraction — clean Markdown & structured JSON for Claude, Cursor & any MCP client. 1,000 free one-time credits, local-Ollama LLM support.

OpenZiti MCP Gateway

openziti

NetFoundry MCP Gateway enables secure, isolated access to Model Context Protocol (MCP) tools across distributed systems without exposing public endpoints. Zero trust gateway for MCP servers. Aggregate, filter, and secure