ERNIE Image

Product Overview: ERNIE Image is Baidu’s open-source text-to-image AI model built on an 8B-parameter Diffusion Transformer, designed to generate images with precise text rendering, structured layouts, and complex multi-object prompts. It addresses the challenge of producing clear, readable text and consistent layouts in AI-generated images, a common limitation in many diffusion models.

Key Features:

• Generates sharp, readable text inside images, ideal for posters, infographics, and UI designs
• Maintains structured layouts across multi-panel designs and complex scenes
• Supports both photorealistic and stylized image generation without mode switching
• Runs locally on a single consumer GPU (24GB VRAM) with no API or cloud costs
• Includes a Prompt Enhancer to automatically expand short prompts for better output quality
• Flexible pricing with one-time credit packs: Starter ($9.9), Pro ($29.9), and Scale ($49.9), all including commercial…