MossImage
Hosted playground for the MOSS-Image text-to-image models, with Chinese and English text and editable JSON layout control.
- AI tool
Overview
MossImage is a hosted playground for the MOSS-Image family of text-to-image models, including the Base and SFT checkpoints. It renders Chinese and English lettering and lets you edit a JSON layout to control where things go in the picture. The site describes the model family as open source and documents running the checkpoints locally with ComfyUI. The license of the hosted site itself is not stated. You get generated images with layout you can control.
Key features
- Hosted MOSS-Image playground
- Base and SFT checkpoints
- Chinese and English lettering
- Editable JSON layout control
- Local ComfyUI instructions
Use cases
Alternatives
Last updated 2026-11-17