18 adjustments, 22 filters, 4 generators. All GPU-ready.
Organized into three menus: Color, Filter, and Generate. Parametrized dialogs give live preview before committing. Every filter is accessible from the Scripting API too.
Automatically stretches the tonal range of the active layer to use the full 0–255 range. One-click instant improvement for many photos.
Converts the layer to grayscale using perceptual luminance weighting (BT.601: 0.299R + 0.587G + 0.114B). Preserves alpha.
Inverts all RGB channels (255 − value). Alpha is untouched. Useful for creating negatives or inverting masks.
Inverts only the alpha channel. RGB values are untouched. Useful for flipping transparency masks.
Applies a classic warm sepia tone. Desaturates and tints toward brown-gold for a vintage photographic look.
Adjust overall luminosity and tonal range. Live preview dialog. Classic linear brightness/contrast controls.
Full per-channel tonal curve editor. Drag control points to remap input/output values. Supports RGB composite and individual R, G, B curves.
Adjusts exposure in EV stops. +1 EV = double the light. Works well for underexposed photos or pulling back highlights.
Independently adjust the bright and dark tonal regions. Recover detail in highlights without affecting shadows and vice versa.
Shift hue globally, boost or drain saturation, and lighten or darken. All in one dialog. GPU-accelerated for instant preview.
Set input black/white points and gamma correction. Clamp and stretch the tonal range with precise numeric or slider control.
Adjust color temperature (warm/cool) and magenta/green tint. Primary controls for white balance correction.
Converts the layer to black and white by applying a hard cutoff. Pixels above threshold become white; below become black.
Reduces the number of distinct tonal levels per channel, creating a bold flat-color poster effect.
Adjusts the balance of color in shadows, midtones, and highlights independently using cyan/red, magenta/green, yellow/blue sliders.
Maps image luminosity to a color gradient. Dark tones map to the first color, bright tones to the last. Great for duotones and stylized looks.
Converts to grayscale with individual per-channel weighting. Adjust how much R, G, or B contributes to the final luminance.
Boosts saturation while protecting already-saturated colors. Primarily affects muted and skin-tone colors, avoiding oversaturation.
Mathematically precise Gaussian blur. GPU-accelerated at large radii. Essential for soft light and depth-of-field effects.
Circular aperture blur simulating optical lens bokeh. More correct for lens-based blur than a simple Gaussian.
Directional blur that simulates camera or subject motion. Set the angle and distance of the blur streak.
Fast, uniform box blur. Less smooth than Gaussian but significantly faster for large radii. Good for glow effects.
Radial zoom blur creating the impression of fast forward motion toward a focal point. Adjustable strength and focal area.
Unsharp mask sharpening. Enhances edge contrast to restore apparent detail lost to softness or compression.
Smooths random noise while attempting to preserve edges. Useful for cleaning up high-ISO photos or scanner grain.
Replaces each pixel with the median of its neighborhood. Extremely effective at removing salt-and-pepper noise without blurring edges.
Groups pixels into Voronoi cells and fills each with its average color. Creates a stained-glass or crystal shard effect.
Perlin-noise based dent/bump texture warp. Creates an irregular, organic surface distortion effect.
Mosaic/pixelation effect. Divides the image into square blocks and fills each with its average color.
Spherical radial distortion. Positive values push pixels outward (fisheye); negative values pull inward (pinch).
Rotational swirl distortion from the image center. Rotation increases with distance, creating a characteristic spiral.
Adds Gaussian (film-like) noise. Monochrome mode adds equal noise to all channels (gray grain); color mode adds independent per-channel noise.
Soft bloom/glow effect. Blurs bright areas and blends them back. Great for dreamy, ethereal looks or neon lighting.
Darkens the edges of the image, drawing the viewer's eye toward the center. Adjust strength and softness.
Renders the image through a regular grid of dots, simulating the classic halftone printing process.
Edge detection + line art rendering that transforms a photo into an ink drawing. Adjustable strength and threshold.
Mode-based smoothing that simulates the blended oil paint look. Larger radius produces a more impressionistic result.
Tints the image with a selected color using multiply/overlay blend. Like placing a colored gel over a photo.
Randomly drags horizontal rows of pixels sideways, simulating corrupted digital signal or tape glitch artifacts.
Offsets R, G, and B channels independently, creating chromatic aberration or vintage VHS color fringing effects.
Remove Background
Cut subjects from their backgrounds with a single click. PaintFE uses deep learning to trace exact silhouettes — including fine hair, transparent glass, and complex edges — that would take minutes of manual masking.
The model runs entirely on your own hardware. No cloud upload. No account. No subscription. Your images never leave your machine. You download the model checkpoint once and it’s yours.
Renders a configurable grid of lines over or below the layer stack. Control line color, spacing, line weight, and opacity.
Generates a drop shadow beneath the active layer’s content based on its alpha channel. Non-destructive; placed on a new layer below.
Creates an outline/stroke around the edges of an alpha-channeled layer. Choose outline size, color, and rendering position.
Traces luminosity or alpha contour lines across the image, creating an isoline topographic map effect or illustrative line-art style.
Every filter is scriptable
All filters listed on this page are available as functions in the
Scripting API: call apply_blur(3.0),
apply_vignette(0.5, 0.3), apply_hsl(30.0, 20.0, 0.0)
and so on. Chain them, loop them, run them on folders via CLI.