Where libvips (pyvips) shines
libvips is a fast, demand-driven image processing library with low memory needs, licensed LGPL-2.1-or-later and used from Python through pyvips. With around 300 operations and wide format support, including JPEG XL, JPEG 2000, HEIC, AVIF, and PDF, it's a favorite for high-volume servers and thumbnailing.
Do it with BrushCue
BrushCue is aimed at creative editing rather than throughput. It runs on the GPU, supports custom WGSL shaders, and has a browser tool for every effect. Side by side:
pyvips
import pyvips
image = pyvips.Image.new_from_file("photo.jpg")
image.gaussblur(8).write_to_file("photo-blurred.png")BrushCue (Python)
import brushcue as bc
ctx = bc.Context()
image = bc.Composition.load("photo.jpg")
# Gaussian blur with a sigma of 8 pixels
edited = image.gaussian_blur(8.0)
result = edited.execute(ctx)
with open("photo-edited.png", "wb") as f:
f.write(result.to_image_bytes(ctx))For resizing and thumbnailing at scale, pyvips is one of the fastest options available. Use BrushCue when the work is creative, such as film looks, distortions, or halftones.
BrushCue Script (bcue)
image = Composition::from_path("photo.jpg")
edited = image.gaussian_blur(8.0)
Bytes::save_to_path(Image::to_bytes(Composition::render_to_image(edited)), "photo-edited.png")



