v1.0.1

Cubric Vision v1.0.1 - First public release🎉

The first official public release of Cubric Vision a local, open-source
(AGPL-3.0) image and video generator powered by ComfyUI that runs entirely on
your own machine. No installer, no admin rights, no cloud account.

Highlights

• Three platforms. Portable builds for Windows (x64), Linux (x64), and macOS
  (Apple Silicon / arm64). Each runs in place — unzip and launch.
• Zero-setup engine. The ComfyUI engine bootstraps on first run (bundled uv;
  resumable model + dependency downloads). No manual Python/ComfyUI setup.
• In-place updater. Online and offline updaters, both validated on hardware.
  Updates preserve your engine, models, projects, and settings.
• No-terminal launch on Windows & Linux. Launches with no console window; a
  *-with-terminal variant is included for watching the live log.

Fixes since the pre-release cycle

• Subfolder LoRA / upscale models now load on Windows. Models in a subfolder
  (e.g. loras/SDXL/foo.safetensors) failed with "Prompt outputs failed
  validation" on Windows; the app now emits the engine-native path separator and
  heals existing projects automatically.

macOS first launch

Builds are not signed or notarized. Clear the download quarantine once in Terminal:

xattr -dr com.apple.quarantine "/path/to/CubricVision-macos-arm64-v1.0.1"

then double-click start.command.

Platform support status

• Windows: tested on the maintainer's Windows dev machine. Not yet validated
  on a separate clean Windows host.
• Linux: install + launch validated on a low-power Ubuntu laptop. Generation
  is unvalidated pending a stronger Linux host or contributor report.
• macOS: validated on a rented Apple Silicon Mac (16 GB unified memory) —
  install, launch, and image generation all confirmed working. Video
  generation was not tested (it needs more memory than the test machine had);
  community validation welcome.

Cubric Vision is a local image and video generation app. Report validation
results or issues via the repo's issue templates.