LLaDA-Image
Online AI tool for Chinese-English text, text-to-image generation, and reference-image instruction editing
About LLaDA-Image
LLaDA-Image is an online AI image-generation and editing workspace using one unified diffusion stack and checkpoint for text-to-image generation, reference fidelity, and instruction-based editing. Upload a reference image and describe changes to update its style or content while keeping the subject consistent, or generate an image directly from a prompt. It suits poster and layout rendering for Chinese-English creative briefs. Text-to-image dimensions must be divisible by 16; editing dimensions by 32.
Best for
Best use cases
Generate images from prompts
Upload a reference image and update its style or content by instruction
Create Chinese-English posters and layouts
Try LLaDA-Image in a browser
Features
Unified generation and editing
One checkpoint covers text-to-image generation, reference fidelity, and instruction-based editing.
Instruction-based reference editing
Upload a reference image and describe changes to update its style or content while keeping the subject consistent.
Prompt-based generation
Generate images with prompts alone. Base and Turbo height and width must be divisible by 16.
Chinese-English text and layouts
Supports poster and layout rendering with Chinese and English text.
Pros
- One unified diffusion stack and checkpoint cover generation and editing
- Reference-image editing updates style or content by instruction while keeping the subject consistent
- Supports poster and layout rendering for Chinese-English creative briefs
Cons
- Text-to-image dimensions must be divisible by 16, and editing dimensions by 32