What M-art is
M-art is a tool for creating images with artificial intelligence. Describe what you want to see, or upload your own reference images — the model draws a new picture from that.
What sets it apart from ordinary generators is that it remembers the whole recipe for every image: your prompt, the mode used, the format and the references. You can always come back and see how a picture came about.
When to use it
- Illustration for an article or a deck — you need an image you can't buy anywhere.
- Variations on one idea — generate ten takes on the same scene and pick the best.
- Working from a reference — you have a photo or an image whose style you want applied to your own subject.
- Visual experiments — see what the model does with an unusual prompt.
How it works (short version)
- Prompt — describe what you want. Any language works; the model handles translation itself.
- Mode — choose whether it draws from text only or also from uploaded images. Details in Three generation modes.
- Format and quality — square, landscape or portrait; three quality levels.
- Result — the image is saved into a scene that you name. You don't have to name it right away — an unnamed scene waits among the unsorted ones.
What it costs
Every generated image is paid for in credits. The current price is in the AI operations price list under "M-art image generation".
If generation fails for any reason, credits are refunded automatically.
Higher quality and larger formats take the model more work, but the credit price is currently flat.
What M-art does not do
- It doesn't edit existing photos like a classic photo editor — it always draws a new image, even when working from a reference.
- It doesn't render text reliably. Lettering inside images tends to come out garbled; plan for that and add text in another tool.
- It can't guarantee the same character across images. When you need a recurring motif, use a finished image as the reference for the next one.