MOSS Image for bilingual lettering and layouts

MOSS Image is an independent browser generator that uses the MOSS-Image model family. Its generator takes a text prompt and returns a new raster image. The most useful starting point is a brief that spells out both the scene and the words that should appear in it.

Write the lettering into the brief

The publisher describes support for English and Chinese lettering. For a menu concept, include the exact names of the dishes or drinks and say which line is the headline. Describe where smaller labels belong and check the spelling and punctuation in the resulting image. If the wording must survive later revisions, keep the approved text separately so it can be typeset over the finished visual.

Describe positions for a more structured composition

The site documents experimental JSON prompts for positioning text and objects. Use a simple arrangement first, such as a heading at the top and a product below. Compare the generated result against those positions before adding more elements. The structured input guides the composition, so inspect the placement and spacing in the actual image. If you need to move a single label after generation, use the result as a reference for a separate editing workflow.

Pick the canvas and verify the output

The generator lists square, portrait and landscape aspect ratios and supports JPEG, PNG and WebP downloads. A portrait canvas suits a poster mockup while a landscape canvas gives a banner headline more room. Review the image at its intended display size before using it. The examples on the site are credited to Sogni and illustrate the underlying model’s capabilities. They were produced outside this independent generator.