StarNet Software

AI-powered software for star removal and noise reduction in astrophotography.

Command-line Tools

This page describes the standalone StarNet and DeepSNR command-line packages. Use these files for command-line workflows, automation, host applications, and standalone processing outside PixInsight.

Command-line tool releases

Command-line archives are for standalone processing, automation, and applications that call StarNet or DeepSNR executables directly. For example, Siril's StarNet integration uses the command-line StarNet installation path.

Release matrix

Current command-line packages are published for Linux x64, Windows x64, macOS x64, and macOS ARM64. StarNet2 and DeepSNR use the same version lines as the current PixInsight modules, but the command-line archives are separate files intended for terminal use, scripts, and host applications that call external executables.

Current ORT/CoreML command-line packages are self-contained. Users do not install a separate TensorFlow, Torch, or backend-runtime package.

Current command-line release matrix

Operating system Architecture StarNet2 version DeepSNR version Backend
Linux x64 2.5.0 1.2.0 ONNX/ORT
Windows x64 2.5.0 1.2.0 ONNX/ORT
macOS x64 2.5.0 1.2.0 ONNX/ORT
macOS ARM64 2.5.0 1.2.0 CoreML

Updates in the latest releases

Recent command-line releases focus on current runtime backends, clearer input/output behavior, and StarNet2 workflow additions:

  • StarNet2 2.5.0 and DeepSNR 1.2.0 use ORT/CoreML as the current public command-line builds; TensorFlow and Torch builds are not part of this release and likely will not be part of future releases.
  • Command-line input handling now preserves source bit depth for supported 8-bit and 16-bit integer images instead of relying on OpenCV's default 8-bit conversion path.
  • Unsupported inputs, including 32-bit floating-point images and unsupported channel counts, are rejected. Convert them to 8-bit or 16-bit grayscale/RGB before running the command-line tool.
  • TIFF and PNG outputs are written as 16-bit images by default; use --eight only when an 8-bit output file is intentionally needed.
  • StarNet2 adds --unscreen for an optional star-layer output alongside the existing starless output and optional subtractive mask.
  • macOS ARM64 command-line packages use native CoreML and are the recommended Apple Silicon lane for best performance.

Command-line download pages