 button at the top right of the Filters Panel below the Navigation.

The Crop panel opens and shows the different options to change the file dimensions of your image.

Select from different aspect ratios provided in the dropdown menu or set your own custom crop.

Hold the [SHIFT] key and drag the arrow to snap images by 15 degree increments for more precise cropping.

Click on the Lock symbol to unlock the ratio size. You can type your specific width or height or resize the image manually.

Use the crop corners and lines to create a freehand crop.

Change the rotation of your image by clicking anywhere around your image in the preview and drag it around. You can review the degree of the angle rotation located in the Crop panel.

To undo your custom crop, click on the "Reset" button to return the crop to its original ratio.

Change the orientation of your crop by selecting the opposing arrow symbol between the Width and Height.

Once you have decided on your crop, click "Apply".

Crop

Auto mode decides and suggests settings to enhance your photo by selecting an AI Model and/or Setting enhancements for you.

In the "Select a model" section, click on the thunderbolt symbol. The Auto setting will share a "Recommended" model based on the image you've imported. 

You can also click into the model name and select Auto settings to further refine details. The Lime green dots indicate which setting has Auto applied.

You can also find the "Auto" button by hovering over the Model name and it appears on the left side of the model panel. 

Turn off AUTO by clicking on the Thunderbolt symbol or clicking anywhere on the AI Model or Settings to change the enhancement values.

Auto Mode

Utilize the Compare Tool to review different AI models at a glance.

Locate the Compare tool next to the AI Models title name:

You can also locate the Compare tool when you hover over a model name. 

A small pop up panel prompts you to select up to four different AI models to compare. You do not need to select four models every time. 2 or 3 models still make the Compare Tool accessible. 

A four window view shows the AI models selected. A blue outline highlights the active quadrant. 

You can click on a window and then change the AI model using the panel on the right.

To add another window to compare, select the 

Once you are satisfied with your comparison, click on 

Compare 

The Capture tool screenshots your before and after images.

Click on the Camera icon Next to "Capture" to take the screenshot. 

This symbol appears when Split View or Side by Side view is in use.

Click on "View in Finder" for MAC or Click on "View in File Explorer" for Windows. 

 Folder on your computer shows where the screenshot is saved. 

The same Capture tool is available with the Compare tool

Capture

Review and see changes you have made to images with the File List located at the bottom of the Gigapixel AI screen.

There are two ways to expand or collapse the file list. 

Hover on the black line that separates the Preview and the File List. The line changes to blue. Click and drag the file list to the desired location across the preview. 

Locate the arrow button on the left side of the "Add Image" button. 

Click on the arrow to collapse the file list. Use the same button to expand the file list. The default shows four files expanded in view. 

You can also choose the fixed expander button in the Gigapixel Main Menu, select 

Use the file list to compare and review what settings you have applied.

The file list is broken down into several smaller categories:

 - Files will be shown in Alphabetical order. When additional images are imported after already using the app, these files will be shown at the bottom of the file list.

 The before and after file size after upscaling or cropping - these values will change based on your Resize Mode settings.

 - the multiplier by which your image is upscaled

 Whatever model is used to enhance the file, the name will be displayed in color. Each AI Model has its own designated color.


 The strength of how much noise is suppressed is indicated by a number value.

 The intensity of how much sharpness is added from the image is indicated by a number value.

 When face recovery is turned on, a number value will appear. The intensity of how much Face Recovery strength is used is indicated by a number value.

 "ON" appears when Gamma Correction is applied. "OFF" appears when Gamma Correction is not applied.

When the user interface is in the Theme: Dark mode, a green Thunderbolt symbol indicates when the 'Auto' setting is applied in the Filters Panel. This setting only applies to the AI Models or Settings.

In light mode, the thunderbolt appears black.



Green numbers indicate values from refining Denoise & Sharpen.

File List

Gigapixel's Command Line Interface (CLI) feature, 

available exclusively to Pro License users

, offers advanced functionality for efficient batch processing and integration into automated workflows. Users can leverage this feature to upscale images with precision and speed directly from the command line, ensuring seamless integration with existing software systems and maximizing productivity. 

After install, you should be able to access it from the command line/powershell/terminal by typing in 

 depending on release type) as the command. 

With no arguments, this should print a usage dialog.


The following examples are written with UNIX-style escape characters. Windows users may need to edit these commands to follow CMD/PowerShell formatting.

-m, --model for model. Valid values are specified in json files but should account for common shortenings (e.g., art, cg, and cgi are valid for Art & CGI model)

If there is a short code missing that you tried and it didn't work let us know

Generative models now supported (4/4/2025)

List of AI Models and their Corresponding Aliases

–mv, --model-version for model version. Valid values are based on the UI model versions, so version 2 is for standard, low res, and high fidelity models

--dn/--denoise, --sh/--sharpen, --cm/--compression for the various model options. Accepts values 1-100. 

--fr, --face-recovery for both enabling and setting face recovery strength. Accepts values 1-100.

--scale, --width, --height for setting upscale type/value. All mutually exclusive.

--res, --resolution for setting pixel density

Valid values are stuff like 300ppi, 150ppcm

-i or --input specifies which files or folders to process.

-r, --recursive should recurse into subdirectories when finding input files

--cf, --create-folder will create the output folder if it doesn't exist

--prefix adds a prefix to the output file name

--suffix adds a suffix to the output file name

--flatten will flatten folder structure if using recursive mode

e.g., input/a/1.png and input/b/2.png would be put in output folder without the a/b directories

-f, --image-format specifies the output file type

Accepts jpg, jpeg, tif, tiff, png, and preserve (default)

jpg/tif vs jpeg/tiff will allow 3 vs 4 character output extensions for flexibility

--tc, --tiff-compression sets the tiff compression type

Valid values are none, zip (default), and lzw

Only used if output type is tiff (either set directly or through preserve)

--pc, --png-compression sets compression level for png outputs

Only used if output type is png (either set directly or through preserve)

--bd, --bit-depth sets the bit depth of the output

--jq, --jpeg-quality sets output jpeg quality

--cs, --colorspace sets what color space to use for output

Valid values are preserve (default), sRGB, Pro Photo, Apple, Adobe, Wide, and CMYK

--icc specifies a custom color profile to use for output

Overrides --colorspace flag except in the case of CMYK

--q, --quiet turns off all logging (some logs may still leak through though)

-p, --parallel enables reading multiple files at once to save time at the cost of memory

Accepts any positive integer. A value of 1 is identical to normal flow, a value of 10 would load 10 images at once.

Note that the parallel reading is capped at 8GB estimate file size (image size + upscaled image size estimate)

–am, --append-model appends model name and scale to the end of the filename (not implemented yet)

--face-recovery-creativity, --frc for creativity, 0 or 1

New models for -m flag: recovery and redefine. recovery accepts additional --mv flag, either 1 or 2 (default).

New parameter options for Generative models include:

--prompt: Text prompt to pass to redefine model

 --frv 2 for v2 (default) or --frv 1 for version

Example for running Redefine with a prompt:

 executable should be on the path by default after install, but if not you can add it to your path. The default paths should be:

Upscale all files in a folder by 2x using auto settings, preserving all aspects of image format (extension, bit depth, etc):

Upscale all files inside input directory recursively

Upscale a single raw and convert it to a jpg without using autopilot (all model parameters are set)

Functions

Crop

Auto Mode

Compare

Capture

File List

Command Line Interface