ComfyUI Impact Pack 🧩 Workflow Open source

What it is

ComfyUI Impact Pack by ltdrdata (GPL-3.0) provides detection-and-redraw tools: detectors find regions such as faces, hands or objects, and Detailer nodes regenerate those regions at higher resolution for sharper results. It also adds SAM segmentation, iterative upscaling, wildcard prompts and many utility nodes, and it is a common dependency in portrait, product and illustration workflows. Its author also created ComfyUI-Manager, and the pack has been part of the ecosystem since ComfyUI's early days, so a large body of tutorials and example graphs uses it.

Key features

• FaceDetailer and generic Detailer nodes to redraw small regions
• Bounding-box and segmentation detectors (with the Impact Subpack for YOLO models)
• SAM integration for precise masks
• Iterative upscaling and tiled processing
• Wildcard and prompt-variation support
• Switches and pipe nodes for large graphs

How to install

In ComfyUI/custom_nodes run git clone https://github.com/ltdrdata/ComfyUI-Impact-Pack comfyui-impact-pack, then pip install -r requirements.txt inside it; or install via ComfyUI-Manager. Some detectors need the separate Impact Subpack.

Pricing and rights

Free, GPL-3.0. Detection and segmentation models have their own licences, so check them before commercial use; the pack itself adds no watermark or limits.

Who it is for

Portrait, character and product creators who need clean faces, hands and small details at high resolution, and anyone upscaling images for print or large displays.

Verdict

A mature, very widely used pack for fixing details. It is complex, and recent pushes (April 2026) are less frequent than in its peak years.