This is a lightweight optimizer that takes your image generation prompts and refines them for better results across Midjourney, Sora, and DALL-E. The server runs remotely via SSE transport, so you don't need to spin anything up locally. You'd reach for this when you're iterating on visual prompts inside Claude and want help expanding vague ideas into detailed, platform-optimized descriptions. The source doesn't detail specific operations, but the core use case is clear: feed it a rough prompt concept and get back something more likely to produce good results with image generators. Useful if you're doing a lot of AI art generation and want Claude to handle the prompt engineering heavy lifting.
claude mcp add --transport sse io.github.evozim-prompt-refiner https://prompt-refiner-mcp.vercel.app/api/mcp