stemdeckapp/stemdeck

[!IMPORTANT]
macOS first launch (no code signing yet). After dragging StemDeck to Applications, clear the Gatekeeper quarantine flag or macOS will say the app is damaged:

xattr -dr com.apple.quarantine /Applications/StemDeck.app

Optional: start from a clean slate. To reproduce a true first-run experience, open each path in Finder via the Go menu, then Go to Folder (Shift+Cmd+G), and move the folders to Trash:

• ~/Library/Application Support/StemDeck
• ~/Library/WebKit/app.stemdeck.desktop
• ~/Library/Caches/stemdeck
• ~/Library/Caches/app.stemdeck.desktop

You can also delete ~/Library/Preferences/app.stemdeck.desktop.plist the same way. This is optional; the app will work without it.

What's new in 0.8.0 Alpha 8

NVIDIA build no longer silently runs on CPU

On some Windows machines the NVIDIA build separated stems on the CPU even though a supported NVIDIA GPU was present. It worked, but far slower, with nothing on screen explaining why. Fixed:

• Self-healing installs. Installing the CPU build and later the NVIDIA build on the same machine could leave a stale "CPU only" marker behind that pinned every future run to the CPU. The NVIDIA build now detects and removes that stale marker on launch, then probes the GPU normally.
• Automatic recovery from a bad first run. A one-time hiccup during setup (a slow driver query, a dropped download) used to pin the CPU permanently with no way back. The app now records why it chose a device and re-probes on the next launch when the previous result came from a failure, so simply relaunching heals it, and adding a GPU later is picked up automatically.
• More reliable GPU detection. nvidia-smi is now found on DCH driver installs that place it only under the driver store, and the first detection is given more time so laptops waking a sleeping GPU are not missed.
• Every device decision is written to logs/setup.log, so a CPU fallback always leaves a trace to diagnose.

Choose your compute device (self-hosted)

Self-hosted and Docker deployments now have a Compute device selector under Settings -> Advanced. Leave it on Auto to use the best available device, or force CUDA / MPS / CPU. Devices that are not present on the machine are shown greyed out, and forcing an unavailable device is rejected instead of silently falling back. The choice applies to the next separation without a restart.

Fixes

• Docker image tags (:edge, versioned, :latest) are published again; a workflow trigger regression had stopped them from updating on release.
• The Settings panel scrollbar no longer overlaps the right-aligned controls in the Advanced tab.

Installing

• macOS: drop the .app into Applications and launch (run the xattr command above first).
• Windows: unzip the downloaded .zip, then run StemDeck.exe from the extracted folder.
• Linux: download the .tar.gz for your hardware, extract it, and run ./StemDeck. Install the WebKitGTK + GTK runtime prerequisites first (FFmpeg is fetched automatically on first launch):

  sudo apt install libwebkit2gtk-4.1-0 libgtk-3-0
  

  The NVIDIA build additionally needs a working NVIDIA driver such that nvidia-smi reports your GPU (the CUDA runtime itself is bundled -- no separate CUDA toolkit install needed). If you have no NVIDIA GPU, use the CPU-only tarball.
• Docker / Unraid: install "StemDeck" from Unraid Community Applications, or pull ghcr.io/stemdeckapp/stemdeck:edge and map port 8000 plus the /app/jobs and /cache volumes. Add --runtime=nvidia for GPU acceleration. See the README for the full docker run command.

Artifact scan

• Windows portable packages (CPU + NVIDIA) scanned with ClamAV in CI before upload.
• Linux portable packages (CPU + NVIDIA) scanned with ClamAV in CI before upload.

Artifact build

• macOS arm64 and x64 DMGs and runtime packs built and inspected on a macOS runner before upload.
• Windows portable ZIPs (CPU + NVIDIA) built on a Windows runner.
• Linux portable tarballs (CPU + NVIDIA) built on a Linux runner.