ByteDanceThis text-guided image editor performs precise, instruction-based changes while preserving the original photo’s structure and details. It understands natural language prompts to adjust lighting and color, remove or add objects, and apply style transfers without harming image integrity. Built on a diffusion transformer with a VAE, it delivers high-quality, consistent edits suitable for creative work and professional retouching. Clear, specific prompts work best, and stepwise editing improves reliability for complex tasks. It supports common image formats and high-resolution outputs, making it useful for product photos, marketing materials, digital art, and restorations. Note: extremely complex, multi-step edits in one prompt may reduce consistency.

