This MCP tool server enables image generation and editing via OpenAI's GPT-4o/gpt-image-1 image APIs within the Model Context Protocol. It supports generating images from prompts with advanced options such as size, quality, and background, and can edit or extend images using a prompt with an optional mask, accepting input as file paths or base64. Outputs can be saved directly to disk or returned as base64, with automatic file-output when payloads exceed MCP size limits. The server is designed to work with Claude Desktop, Cursor, VSCode, Windsurf, and any MCP-compatible client. It includes guidance for installation and configuration, including local deployment and Azure deployments, with environment variable support (OPENAI_API_KEY, AZURE_OPENAI_API_KEY, AZURE_OPENAI_ENDPOINT, OPENAI_API_VERSION) and optional env-file usage. Development and run instructions are provided (TypeScript source at src/index.ts, build with yarn, run with node dist/index.js). File paths must be absolute, and a default file location is /tmp or a directory specified by MCP_HF_WORK_DIR. The server relies on the MCP SDK and OpenAI Node.js SDK for image API access.