Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
MOSS-Image
MOSS-Image is OpenMOSS's text-to-image family with a 5B backbone, 4K-token prompts and JSON layout control for placing objects, text and boxes.
AI Tools
Explore AI tools
Free AI Image
Truly Free AI Image Generator
AI Image Editor
AI Photo Editor Online
Image Upscaler
Enhance image resolution effortlessly
AI Color Changer
Transform photo colors in seconds
AI Room Designer
Design your dream room with AI-powered 3D rendering
Skin Enhancer
Enhance skin texture naturally with AI
Face Editor
Enhance face texture naturally with AI
AI Clothes Changer
What Makes MOSS-Image Different
A 5B text-to-image model family built for precise, text-driven layout control.
- Layout Control Built InAssign every object and text element its own bounding box, so the composition follows your JSON spec rather than guesswork.
- 5B Model, 4K-Token PromptsExtended prompts fit with room to spare, and the same weights can be fine-tuned for your own tasks and domains.
- Base and SFT CheckpointsBoth checkpoints use the same architecture and ship under Apache-2.0; the SFT variant produces cleaner whiteboard text.
How to Use MOSS-Image
Follow three steps, from environment setup to a controlled layout render.
MOSS-Image: Frequently Asked Questions
Answers to common questions about MOSS-Image prompts, checkpoints, hardware and setup.
What can I control with native layout prompts?
A JSON prompt describes the scene, objects and rendered text, with a bounding box for each element's position and size.
How are MOSS-Image-Base and MOSS-Image-SFT different?
They share the same architecture and sampling interface; SFT is fine-tuned on curated data and, in the project example, delivers clearer text grouping and a tidier whiteboard layout.
What is the maximum prompt length for MOSS-Image?
Up to 4K tokens, which is enough for a multi-element layout JSON.
What hardware and software are required?
Python 3.10+ and an NVIDIA GPU with a matching PyTorch build. If memory is limited, switch to Sequential mode or reduce image size; Compile DiT is off by default and unavailable in that mode.
Does MOSS-Image work inside ComfyUI?
Yes. Find MOSS-Image in ComfyUI Manager to install the native plugin, then load the example workflow; integration into the official ComfyUI repository is still in progress.
Which license covers the weights and code?
Model weights are released under Apache-2.0, while the pipeline code and third-party components retain their own licenses. For research use, cite the project repository.
Create with MOSS-Image Today
Install the Diffusers pipeline or ComfyUI plugin, then test a JSON layout prompt — or refine your idea with the MOSS-Image Prompt Enhancer first.