This extension for Forge WebUI enables precise control over image generation by allowing users to define specific regions for different prompts, effectively managing character features and avoiding prompt bleeding. It targets Stable Diffusion users seeking to generate complex scenes with multiple distinct subjects or elements.

How It Works

The extension offers multiple modes: Basic (automatic tiling), Advanced (manual bounding boxes), and Mask (custom drawn regions). Each mode allows assigning specific prompt lines to image areas, with weights to control influence. This approach provides granular control over composition, ensuring distinct elements are rendered as intended by the user's prompt.

Quick Start & Requirements

• Install via the Forge WebUI extension manager.
• Compatible with Forge, reForge, Forge Classic, and Automatic1111 WebUI.
• Supports SD1 and SDXL models.
• Official API documentation available: https://github.com/Haoming02/sd-forge-couple/wiki/API

Highlighted Details

• Supports "Common Prompts" for defining reusable prompt snippets.
• Includes a "Tile Mode" for high-resolution upscaling with per-tile prompting.
• Offers "Global Effect" to apply a prompt to the entire image.
• Advanced and Mask modes allow for precise region definition and manipulation.

Maintenance & Community

• Active development and community support are implied by the detailed README and multiple modes.
• No specific community links (Discord/Slack) or roadmap are provided in the README.

Licensing & Compatibility

• The README does not explicitly state a license. Compatibility with commercial or closed-source projects is undetermined.

Limitations & Caveats

Regions are suggestions, not strict boundaries, and effectiveness depends on the checkpoint model. Hires. fix prompts are not supported. For Automatic1111, manual patch removal may be required if errors occur during generation.