ZenCtrl Demo

[WIP] One Agent to Generate multi-view, diverse-scene, and task-specific high-resolution images from a single subject image—without fine-tuning. We are first releasing some of the task specific weights and will release the codes soon. The goal is to unify all of the visual content generation tasks with a single LLM...

Modes:

• Subject Generation: Focuses on generating detailed subject portraits.
• Background Generation: Creates dynamic, vibrant backgrounds: You can generate part of the image from sketch while keeping part of it as it is.
• Canny: Emphasizes strong edge detection.
• Depth: Produces images with realistic depth and perspective.

For more details, shoot us a message on discord.