P Image Upscale

Fast Pruna AI image upscaler for increasing existing images up to an 8 MP output.

Model Overview

P Image Upscale takes an input image and produces a higher-resolution version. This Nodespell node exposes the practical controls: target megapixels or scale factor, optional realism/detail enhancement, output format, and safety checker handling.

Best At

• Fast image upscaling for existing images that need a cleaner larger output.
• Choosing either a fixed output megapixel target or a side-length scale factor.
• Simple image-in, image-out workflows where the original composition should stay recognizable.

Limitations / Not Good At

• The model caps output at 8 MP, so very large scale factors may be clipped by the vendor.
• Detail and realism enhancement can change texture or contrast instead of only resizing pixels.
• This node is not prompt-driven image generation or inpainting.

Ideal Use Cases

• Preparing small images for larger previews, social assets, or downstream editing.
• Upscaling AI-generated images before compositing or publishing.
• Batch-style graph workflows that need predictable image resizing with optional enhancement.

Input & Output Format

• Input: required image, plus upscale_mode, either target or factor, optional enhancement toggles, and output settings.
• Output: upscaled image returned on response.

Performance Notes

• Replicate prices this model by output image tier: up to 4 MP or above 4 MP.
• Nodespell estimates from the selected target or from input megapixels multiplied by the scale factor, then settles using the measured output image size.