image-resize — Traditional AI Model
Resize to an exact width and height, from one pixel to 16384, with Bicubic, Bilinear or Lanczos interpolation and optional aspect lock. When you want a tier rather than a number, use Conditional Resize. CPU, instant.
Details
- CategoryTraditional
- Year-1
- License
Inputs & Outputs
- ImageInputimage
Source image to resize
- ImageOutputimage
Resized output image
Parameters
- methodstringdefault Bicubicoptions: Bicubic, Bilinear, Lanczos
Interpolation method
Resampling algorithm. 'Bicubic' (default) gives the best balance for general photographic content. 'Bilinear' is fastest but slightly softer. 'Lanczos' produces the sharpest result but can introduce ringing on hard edges — best for upscaling, riskier for downscaling.
- widthintegerdefault 512range 1–16384
Output width (px)
Target output width in pixels (1-16384). When keep_aspect_ratio=true, this is treated as the maximum width and the height scales proportionally to fit.
- heightintegerdefault 512range 1–16384
Output height (px)
Target output height in pixels (1-16384). When keep_aspect_ratio=true, this is treated as the maximum height and the width scales proportionally to fit.
- keep_aspect_ratiobooleandefault true
Preserve aspect ratio
When true (default), the image is fit into width×height without stretching — the longer dimension hits its target and the other scales proportionally (may produce a smaller image). When false, the image is forced to exactly width×height (may stretch/squash).
Tags
Alternatives in Image Resizing
- Image Conditional Resize
Auto-resize image to the next standard resolution tier.
- Image Resize To Match
Resize image to match another image's dimensions.