AI image model
MOSS Image Prompt Guide
Practical tips and ready-to-use prompts for MOSS Image, from bilingual posters to JSON layouts.
Unofficial: TrendFX is an independent site, not affiliated with or endorsed by OpenMOSS. MOSS Image is an open-source model by OpenMOSS (Apache-2.0). Official sources: openmoss.ai and the OpenMOSS GitHub.
Prompt tips
- Quote visible text. Wrap every word that should appear in "quotes", exactly as written.
- Name the format. Poster, label, sign, banner, plus an aspect ratio.
- Place things. Top, center, bottom, left or right, or use a JSON layout.
- Keep text short. Fewer words usually render more accurately.
- Describe style last. Colors, lighting and illustration style after the content.
The fal API for MOSS Image accepts natural language, tag sequences and JSON layouts, and has an optional prompt expansion step that plans a layout while keeping your quoted text.
10 example prompts
- Bilingual café poster
A minimalist café poster. Headline "OPEN LATE" in bold sans-serif at the top, Chinese subtitle "深夜咖啡" below it, a steaming cup illustration, warm cream background.Use this prompt - Shop sign photo
Photo of a small bakery storefront at dusk, wooden sign above the door reading "Corner Bakery 街角面包店", warm window light.Use this prompt - Event flyer
Event flyer, 9:16. Title "BOOK FAIR 2026", Chinese line "书香满城", footer "Free entry · All ages". Flat illustration of stacked books.Use this prompt - Sticky note
Close-up of a yellow sticky note on a fridge, handwritten in Chinese: "记得买牛奶和鸡蛋!"Use this prompt - Product label
Studio shot of a glass jar of honey with a paper label reading "Wild Flower Honey" and "野花蜂蜜 500g", soft daylight.Use this prompt - Menu board
Chalkboard menu on a café wall: "TODAY" at the top, then "Latte 28", "Matcha 抹茶 32", "Croissant 18".Use this prompt - Social quote card
Square quote card, pastel peach background, centered text "Small steps every day" and below "每天进步一点点", simple line flower.Use this prompt - Workshop banner
16:9 banner for a lettering workshop. Left: headline "Hand Lettering Workshop" and "手写字体工作坊". Right: brush and ink illustration.Use this prompt - JSON layout
{"canvas":"poster, 3:4","background":"light butter yellow","elements":[{"type":"text","content":"GRAND OPENING","position":"top center","style":"bold"},{"type":"text","content":"开业大吉","position":"center"},{"type":"image","content":"red ribbon and scissors","position":"bottom"}]}Use this prompt - Street banner
Photo of a festival street banner strung between two lamp posts reading "Lantern Festival 元宵节快乐", evening, paper lanterns.Use this prompt
Aspect ratio:
Cost: 2 credits (free in demo)Your image will appear here.
Planned pricing: $9.9 for 200 credits, 1 image = 2 credits (about 100 images). Payments are not live yet; generation is free in demo mode.
Frequently asked questions
Is this the official MOSS Image website?
No. TrendFX is unofficial and not affiliated with OpenMOSS. MOSS Image is released by OpenMOSS; see openmoss.ai and their GitHub for the official model.
What prompt format does MOSS Image accept?
Natural language, tag sequences and JSON layouts.
How do I make sure text appears exactly?
Put the text in double quotes, keep it short and say where it should go. Always proofread the result.
What is a JSON layout prompt?
A structured prompt that lists elements such as text and objects with their positions, so the model can follow a layout.
Should I write prompts in English or Chinese?
Either works. Quote the visible text in the language you want it to appear in.