# Patchy: an open-source image editor that keeps PSD layers editable

> Patchy is a free, open-source image editor for Windows, macOS, Linux and the browser, built around PSD compatibility so text, vectors, masks, layer styles and Smart Objects stay editable. It ships signed desktop builds plus a Flatpak and an AppImage for Linux, and the maintainer publishes release notes and SHA-256 checksums for every build.

**SethRobinson/Patchy** — Open-source image editor focused on PSD compatibility and workflows familiar to Adobe Photoshop users.

- Repository: https://github.com/SethRobinson/Patchy
- Stars: 619 · Forks: 37
- Language: C++
- License: MIT
- Published: 2026-10-22 · Updated: 2026-10-22 · Language: en
- Canonical page: https://hysenlabs.com/projects/sethrobinson-patchy

## What Patchy is trying to be

Patchy is a free, open-source image editor for Windows, macOS, Linux, and the browser. The project describes itself as built with a focus on accurate PSD compatibility, keeping text, vectors, masks, layer styles and Smart Objects editable when working with layered Photoshop files.

That is the whole pitch, and it is a more specific claim than open source image editor. Most of the free alternatives in this space are raster editors that import a PSD, flatten what they cannot represent, and hand the rest back as a single layer. Patchy is explicitly built against the case where the file has to survive a round trip, and the list of what it keeps editable is long: text as text, vectors as vectors, masks, layer styles, and Smart Objects.

Two other capabilities sit alongside that. You can work in a limited palette without flattening your layers, which is the kind of constraint that breaks naive editors. And you can import editable Affinity and PDF artwork, so a file that never was a PSD can still arrive with editable structure.

The interface is deliberately familiar to anyone who has used Photoshop, down to the keyboard shortcuts, and the project says so directly: designed to feel familiar if you are used to Photoshop's workflows and keyboard shortcuts. That is a design decision with a double edge. It lowers the learning curve for a huge existing user base, and it also means the application is judged against a mature product's expectations.

It also extends with classic plug-ins and JavaScript, and the repository has an agent-kit/ directory and a patchy-mcp connector, so the project is being built to be driven by an AI agent as well as by a person.

## Download packages and how the signing differs per platform

Every release is published on the GitHub Releases page with SHA-256 checksums, and the current release listed in the documentation is 1.14 from October 19, 2026.

The packaging table covers six artefacts. On Windows 10 and 11 64-bit there is an installer, PatchyWindowsInstaller.exe at 100 MB, and a ZIP with no installer at the same 100 MB. On macOS 12 and later on Apple Silicon there is a DMG to drag into Applications, PatchyMacOS.dmg at 111 MB. On Linux x86_64 there is a Flatpak from the project repository, a single-file Flatpak bundle at 68 MB, and an AppImage for Ubuntu 22.04 and later at 119 MB. And for any modern browser there is nothing to install, since the editor runs at patchyimageeditor.com.

The signing story is worth reading carefully, because it differs by platform. Windows releases are code signed by Seth A. Robinson. The macOS app is signed and notarized by Robinson Technologies Corporation. The Linux Flatpak repository is GPG signed. Three different trust chains for one application, which is normal for a project that ships on three operating systems and explains why a Windows user and a Linux user verify in completely different ways.

## Installing on Linux with flatpak or the AppImage

The Flatpak route installs for your user and pulls the shared KDE runtime from Flathub, and it needs no root:

```bash
flatpak install --user -y https://rtsoft.com/flatpak/com.rtsoft.patchy.flatpakref
```

After that, Patchy updates with flatpak update or through your software center, and starts from the application menu or with flatpak run com.rtsoft.patchy. The flatpak tool itself has to be present, and the documentation says to install it first with sudo apt install flatpak if needed.

There is a caveat about the software centre path that is worth repeating. The .flatpakref link in the download table downloads a small file that a Flatpak-aware software centre such as GNOME Software or KDE Discover can open to install Patchy, but Ubuntu has none of these by default, and on Ubuntu 24.04 the Open button in GNOME Software fails to start Flatpak apps, so the terminal command is the reliable route.

The AppImage is the alternative for people who would rather download a single file. It runs without installing anything on Ubuntu 22.04 or newer, or another distribution from 2022 on, and it updates itself:

```bash
mkdir -p ~/Applications && cd ~/Applications
wget -O PatchyLinux.AppImage https://github.com/SethRobinson/Patchy/releases/latest/download/PatchyLinux.AppImage
chmod +x PatchyLinux.AppImage
./PatchyLinux.AppImage
```

The AppImage needs FUSE 3, though not libfuse2, and the documentation gives the install command per distribution: sudo apt install fuse3 on Ubuntu, Mint, Debian and Pop!_OS, sudo dnf install fuse3 on Fedora, sudo zypper install fuse3 on openSUSE, and sudo pacman -S fuse3 on Arch and Manjaro. Where FUSE is missing entirely, such as in some containers and minimal systems, you run it with the --appimage-extract-and-run argument instead.

## Getting the AppImage into the menu and driving it from an agent

An AppImage is a file in a folder, and the documentation addresses that with integration tools. Gear Lever is a Flatpak app that opens the AppImage with it, AppImageLauncher offers to integrate an AppImage the first time you open it, AM does the same from the command line with an --launcher argument, and AppImage Manager is the fourth option listed.

Updates work at two levels. When a new version is out, Patchy's own update prompt can download it, replace the AppImage file and start again, provided the file sits in a folder you can write to, and otherwise it offers the download. The AppImage tools themselves, such as Gear Lever or AppImageUpdate, can also perform the update.

The agent control angle is the most unusual item in the documentation. For AI control, the connector command is the AppImage file with the argument patchy-mcp, and the Help menu entry Set up AI Control shows the exact line. So the same binary that opens a window can be invoked as a connector, which is why the repository carries an agent-kit/ directory.

That combination, a graphical editor that also exposes an agent connector, is a reasonable thing to want in an automated pipeline, and it is also a fairly new surface area. The documentation shows the mechanism but not a full workflow, so treat it as a capability to evaluate rather than a settled feature.

## How the project is organised and how it is built

The repository is a C++ project with a CMake build. The top level has CMakeLists.txt and CMakePresets.json for the build system, a src/ directory for the code, a third_party/ directory, a NOTICE-THIRD-PARTY.md file for third party attribution, and a LICENSE file.

Around that there is a lot of supporting structure, which tells you what the maintainer is investing in. There is a packaging/ directory holding the branding assets referenced by the README, a docs/ directory with a screenshots document, an agent-kit/ directory, a scripts/ directory, a translations/ directory, and a themes/ directory. The test side has a test/ directory, a test-fixtures/ directory with fixtures, a testy/ directory, and a latest_version.json file at the root.

The release cadence is fast. Version 1.12 landed on 2026-10-16, 1.13 on 2026-10-17 and 1.14 on 2026-10-19, and the last push was on 2026-10-20, so the tree is a day old. There is also a RELEASE-HISTORY.md file at the root, which is where you would look to understand what changed between versions rather than reading release notes on the releases page.

An AGENTS.md file sits at the top level next to the source, which is a strong hint that the project expects an AI agent to work on it and has written down the rules for doing so.

## Where Patchy is the wrong tool

The first limit is that the documentation says nothing about the parts of a professional image workflow that people actually get paid for. There is no mention of colour management and profiles, of CMYK or spot colour, of print resolution targets, of raw camera formats, or of a plugin ecosystem comparable to Photoshop's. The features it does list are about the file format surviving a round trip, and that is where the effort has gone.

The second is maturity. A 1.14 release three days after 1.12 is a project in active development, which is good for momentum and risky for a professional workflow where a regression in PSD parsing is expensive. The upside of the fast cadence is that gaps do get filled. The downside is that you should pin a version and test against real files.

The third is the browser build. It is convenient, since nothing is installed, but a browser hosted editor is a different proposition from a local application when you are handling client artwork, and the documentation does not say whether the browser and desktop builds share the same code path for every feature.

The fourth is the platform floor on macOS, which is Apple Silicon only, and Linux, which is x86_64 only. The repository topics mention affinity, editor, image, layer, photoshop, psd and styles, but there is no other platform in the download table.

Compare that with the alternative of GIMP, which is a mature raster editor with a large plug-in ecosystem and far more colour management tooling, at the cost of a steeper interface and weaker PSD fidelity. The difference in approach is exactly the point: GIMP is a general raster editor that also imports PSD files, while Patchy is a PSD compatibility layer with an editor attached, so it wins on layered file fidelity and loses on everything outside that.

## Licence, verification and the upgrade path

Patchy is released under the MIT licence, with the text in the LICENSE file at the repository root, and there is a separate NOTICE-THIRD-PARTY.md file for third party attribution, which is what you would expect from a C++ application of this kind that pulls in libraries it does not own.

The MIT terms permit use, modification and redistribution, including inside a commercial product, provided the copyright notice and licence text accompany copies. That is a short obligation, but the third party notices are the part to check carefully before you redistribute a build, because the bundled libraries carry their own terms.

Verification is documented per platform. The GitHub Releases page publishes SHA-256 checksums for every release, so you can confirm the file you downloaded matches the one that was published. The Windows build is code signed, the macOS app is signed and notarized, and the Flatpak repository is GPG signed, which means a Flatpak install from the project repository is verified by GPG on the repository side.

For upgrades, there are two paths. The Flatpak route updates with flatpak update or through the software centre. The AppImage route updates through Patchy's own prompt, which downloads the new file, replaces the old one and starts again, provided the file is in a writable folder, and the AppImage integration tools can update it as well.

The browser build needs no upgrade at all, which is a genuine advantage of the hosted option, and the reason to keep it in mind when you decide how you will actually work.

## Conclusion

Patchy fits a designer or developer who works with layered Photoshop files and wants them to stay editable in a free tool, and who can accept a young application that is still filling in the gaps of the Photoshop workflow. It does not fit a professional retouching pipeline that depends on a plugin ecosystem and colour management that the project does not document. Before adopting it, open a representative PSD with Smart Objects and layer styles in the browser build and check what survives, read the release notes for the current version, and confirm that the platform build you want is code signed, since the Windows, macOS and Linux packages are signed in three different ways.

## FAQ

### How do I install Patchy on Linux?

For Flatpak, run flatpak install --user -y https://rtsoft.com/flatpak/com.rtsoft.patchy.flatpakref, which needs no root and pulls the KDE runtime from Flathub. Alternatively download PatchyLinux.AppImage, make it executable and run it; the AppImage needs FUSE 3, or you can use the --appimage-extract-and-run argument where FUSE is unavailable.

### Does Patchy keep PSD layers editable?

The project states that it is built with a focus on accurate PSD compatibility, keeping text, vectors, masks, layer styles and Smart Objects editable when working with layered Photoshop files. It also imports editable Affinity and PDF artwork.

### Which platforms does Patchy run on?

Windows 10 and 11 64-bit as an installer or a ZIP, macOS 12 and later on Apple Silicon as a DMG, Linux x86_64 as a Flatpak, a Flatpak bundle or an AppImage, and any modern browser with nothing to install.

### How can I verify a Patchy download?

Every release on the GitHub Releases page is published with SHA-256 checksums. Windows releases are code signed by Seth A. Robinson, the macOS app is signed and notarized by Robinson Technologies Corporation, and the Linux Flatpak repository is GPG signed.

### Can Patchy be controlled by an AI agent?

The documentation describes an agent connector: the command is the AppImage file with the argument patchy-mcp, and the Help menu entry Set up AI Control shows the exact line to use. The repository also contains an agent-kit directory.

### What licence does Patchy use?

Patchy is released under the MIT licence, with the terms in the LICENSE file at the repository root. The repository also carries a NOTICE-THIRD-PARTY.md file for the licences of bundled third party libraries.

## Sources

- [Issues](https://github.com/SethRobinson/Patchy/issues)
- [License: MIT](https://github.com/SethRobinson/Patchy/blob/main/LICENSE)
- [README](https://github.com/SethRobinson/Patchy/blob/main/README.md)
- [Releases](https://github.com/SethRobinson/Patchy/releases)
- [SethRobinson/Patchy on GitHub](https://github.com/SethRobinson/Patchy)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/sethrobinson-patchy