OpenAI Tools MCP Server

设计与媒体 HTTP 可写/可执行：会让 AI 修改或操作真实资源

能做什么：封装 OpenAI 的图像与音频接口：文生图、改图、生成变体，做语音转写与翻译，另可列模型、查模型和跑内容审核。接口范围聚焦在图像与语音两类能力。

什么时候接：想在 agent 里直接调 OpenAI 的出图和转写能力，又不想自己写一层 HTTP 封装时，接这个最省事。

注意：需 OpenAI API key，按官方价计费；deleteModel 会真删模型记录，createModeration 只给分级结果，不给处置建议。

可用工具

createImageCreates an image given a prompt.
createImageEditCreates an edited or extended image given an original image and a prompt.
createImageVariationCreates a variation of a given image.
createTranscriptionTranscribes audio into the input language.
createTranslationTranslates audio into English.
listModelsLists the currently available models, and provides basic information about each one such as the owner and availability.
retrieveModelRetrieves a model instance, providing basic information about the model such as the owner and permissioning.
deleteModelDelete a fine-tuned model. You must have the Owner role in your organization.
createModerationClassifies if text violates OpenAI's Content Policy

更新日期：2026-10-29（tools/list 实测）

{
  "mcpServers": {
    "remote-ai-com-mcp-openai-tools": {
      "type": "http",
      "url": "https://openai-tools.run.mcp.com.ai/mcp"
    }
  }
}

资料来自官方仓库与 MCP 接口探测，以官方最新文档为准；安装第三方 MCP server 前请自行评估安全风险。