Feedback
freeTrialImage.bannerPity
freeTrialImage.upgradeUnlock
- ✓freeTrialImage.benefitHd
- ✓freeTrialImage.benefitWatermark
- ✓freeTrialImage.benefitUnlimited
MOSS-Image
Meet MOSS-Image: an open text-to-image family that follows JSON layouts, keeps 4K-token prompts in memory, and renders labels exactly where you set them.
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
Where MOSS-Image Excels
An open-weight family from OpenMOSS that treats composition as part of the prompt — MOSS-Image renders what you place, not what it guesses.
- Layout Written Into the PromptEach object and each string of rendered text gets its own bounding box, so the arrangement you describe is the arrangement you receive — no reroll roulette.
- A 5B Core with 4K-Token ContextLong, detailed briefs fit in a single pass, and the published weights can be tuned further for your own niche, dataset or visual domain.
- Two Checkpoints, One LicenseBase and SFT share an identical architecture under Apache-2.0; the SFT build simply keeps dense whiteboard lettering tidier and better grouped.
From Setup to First Render with MOSS-Image
Three quick moves take you from a clean environment to a MOSS-Image layout you fully control.
Questions About MOSS-Image
Answers on layout prompts, checkpoints, hardware requirements and licensing for MOSS-Image.
What exactly can I pin down with a layout prompt?
Your JSON names the scene, its objects and any text to render, then assigns every element a box that fixes where it sits and how large it appears.
How is the SFT build different from the Base build?
Both run on the same architecture and sampling interface. The SFT build was fine-tuned on curated data and, in the project's own example, yields tighter text grouping and a neater whiteboard.
How much text can one prompt hold?
Roughly 4K tokens, which leaves plenty of headroom for a multi-element layout written as JSON.
Which hardware and software are required?
Python 3.10+ plus an NVIDIA GPU running a matching PyTorch build. Short on memory? Use Sequential mode or a smaller image size — Compile DiT stays off by default and is not available there.
Does MOSS-Image work inside ComfyUI?
Yes. Install the native plugin through ComfyUI Manager, then open the bundled example workflow; integration into the official ComfyUI repository is still underway.
What licensing covers the weights and the code?
Weights ship under Apache-2.0, while the pipeline code and third-party pieces keep their own terms. If you publish research, cite the project repository.
Your First MOSS-Image Layout Is One Prompt Away
Spin up the Diffusers pipeline or the ComfyUI plugin, drop in a layout prompt, and watch each box land where you planned — or draft the idea with the built-in prompt helper first.