gen-image MCP server

This is a small helper that lets your AI make and edit pictures for you. You bring your own image service (like an OpenAI-style or Gemini image API), and the finished images land right in your project folder. It is handy for anyone who wants pictures generated while they work, without leaving their chat.

What this MCP server does

You ask your AI for an image, like a red cube on a white desk. Your AI passes that request to this helper. The helper sends it to the image service you configured, using your own address, key, and model. When the picture comes back, the helper saves it into your project folder and shows you a preview. If one model is busy or fails, it can try the next one you listed.

Click to zoom

What you can do with it

• Create a brand new image from a text description
• Edit an existing local image, like changing the sky to a sunset
• Combine or use a local reference image to guide the result
• Save finished images directly into your project folder
• See a preview of the image right in the chat
• List which image models you have set up
• Let it automatically try another model if one fails

Try asking your AI

• “Make me an image of a red cube on a white desk with soft light and save it to exports/cube.png”
• “Edit inputs/photo.png and change the sky to a sunset, keeping the building details”
• “Turn this sketch into a watercolor painting, use a 16:9 shape”
• “Which image models do I have configured right now?”

What it gives back to you

You get back the saved file path and a link to the image, plus a small preview in the chat when the image is not too large. It also tells you which model actually worked, how long it took, and how many tries happened. If something failed, you get a clear error message instead of a picture.

Before you start

Who it's for

People who make things and want images generated or edited on the fly, like developers, designers, and writers working inside an AI chat.