v2.5 is out!

🆙 Upscayl

Free and Open Source AI Image Upscaler

Upscayl lets you enlarge and enhance low-resolution images using advanced AI algorithms. Enlarge images without losing quality, it's almost like magic! 🎩🪄

Upscayl is a cross-platform application built with the Linux-first philosophy.
This means that Linux users receive pre-release builds earlier but Upscayl itself is available on all major desktop operating systems :)

https://upscayl.org

- NOTE: You'll need a Vulkan compatible GPU to upscale images. Many CPU and iGPUs do not work but no harm in trying.

Upscayl should be available on the Software Store of most Linux operating systems. You can download the Flatpak or Snap version, if your distro's Store App supports them.

1. Go to releases section

2. Download the upscayl-x.x.x-linux.AppImage file.

3. Right Click AppImage -> Go to Permissions tab -> Check 'allow file to execute' and then double click the file to run Upscayl.

You can also choose to install using other formats like RPM (Fedora), DEB (Debian/Ubuntu based), and ZIP (Any Linux OS).

(macOS 11 and later)

1. Go to releases section

2. Download the upscayl-x.x.x-mac.dmg file.

3. Double click dmg, drag Upscayl icon into Applications folder.

4. Open Finder, click 'Applications' tab in the left sidebar. Find Upscayl and right click on it. Select 'Open'.

5. In the window that appears, press 'Open' yet again.

(Windows 10 and later)

1. Go to releases section

2. Download the .exe file.

3. Double click exe file, wait for installation, profit.

Check out our Wiki here.

• Try out even more new models!
• Convert your own models
• Compatibility List
• Troubleshooting

Check out Upscayl before/after comparisons here.

You can track all the progress here: https://github.com/orgs/upscayl/projects/1

• Fix bugs
• Make the whole world use FOSS (WIP 🚧)

I recommend using Volta: https://volta.sh for installing Node.js. Download and install volta, then do: volta install node.

# FOR EVERYONE
git clone https://github.com/upscayl/upscayl
cd upscayl

# INSTALL DEPENDENCIES
npm install

# RUN THE DEVELOPMENT SERVER LOCALLY
## YOUR LOGS WILL NOW APPEAR IN THE TERMINAL
npm run dev

# FOR DEVS
## PACKAGE THE APP
npm run dist

## PUBLISH THE APP, MAKE SURE TO ADD GH_TOKEN= IN SHELL
npm run publish-app

• How does Upscayl work?
  • Upscayl uses AI models to enhance your images by guessing what the details could be. It uses Real-ESRGAN to achieve this.
• Is there a CLI available?
  • The CLI tool is called real-esrgan-ncnn-vulkan and it's available on the Real-ESRGAN repository.
• Do I need a GPU for this to work?
  • Yes, unfortunately. NCNN Vulkan requires a Vulkan compatible GPU. Upscayl won't work with most iGPUs or CPUs. But hey, no harm in trying ;)
• Why can't batch upscaling output other formats such as JPGs instead of PNGs?
  • This is due to an issue with REAL-ERSGAN that we can't fix currently tracked here.
• How can I contribute?
  • You can fix code, add features by submitting PRs, report issues or donate :)
• What's the GPU ID for?
  • It is for selecting which GPU to use. The specific procedure is detailed in the Wiki.
    • Note that for Windows systems, if Upscayl is not set the performance mode, the system may override this setting.
• Why does the scale option result in a mosaic/mixed grid?
  • This is due to a bug detailed here and in the Wiki.
• You are not fully open source!!😡
  • We are! We're using completely free and open source tech. We do ship a binary but that's because not including it would kill the entire purpose of the project. If you want to deal with all those PyTorch, NCNN, and Python errors, you're more than welcome to compile your own binary using the code provided by Real-ESRGAN. The binary is pre-compiled for convenience, if you do not trust it, then you can compile your own since everything is free and open source here :)

Upscayl uses Real-ESRGAN-ncnn-vulkan binaries to upscale images. More models and algorithms will come soon.

• Real-ESRGAN for their wonderful research work.
  Real-ESRGAN: Copyright (c) 2021, Xintao Wang
• @JanDeDinoMan, @xanderfrangos, @Fdawgs, @keturn for their code contributions :)
• @aaronliu0130 for providing community support :)
• Foolhardy for their Remacri model.
• Kim2091 for their Ultrasharp and Ultramix Balanced model.