Loom

Experimental preview 0.1.0-preview.8-9a19621 / Standalone tools

PlateSolve
command line.

Find your image’s place in the sky. Run Loom’s blind plate solver from a terminal, independently of PixInsight.

Choose your download.

Choose the operating system and processor architecture of your computer. All four platform packages are preview.8-9a19621. Linux ARM64 uses Vulkan; Jetson Orin hardware qualification remains pending.

Windows x64

Version 0.1.0-preview.8-9a19621 · 14.1 MB

Download for Windows x64

loom-solve-0.1.0-preview.8-9a19621-x86_64-pc-windows-msvc.zip

SHA-256 checksumeec0c95ac8f391d261defc2ffbacd6a0eb8d696346f21fd2bcfc060eab048ff0

14809271 bytes

macOS arm64

Version 0.1.0-preview.8-9a19621 · 13.9 MB

Download for macOS arm64

loom-solve-0.1.0-preview.8-9a19621-aarch64-apple-darwin.zip

SHA-256 checksum7fc5089d9476b1f33db6fcbc56dbfdcf0ee8d55cbbd99d81f18ae620fb9bfa55

14613281 bytes

Linux x64

Version 0.1.0-preview.8-9a19621 · 14.2 MB

Download for Linux x64

loom-solve-0.1.0-preview.8-9a19621-x86_64-unknown-linux-gnu.zip

SHA-256 checksumeb3f035ebf6436a54eadc244952083c33fffc9ac3792a6bacfa57c29ef546c24

14909673 bytes

Linux arm64 · Vulkan

Version 0.1.0-preview.8-9a19621 · 14.3 MB

Download for Linux arm64

loom-solve-0.1.0-preview.8-9a19621-aarch64-unknown-linux-gnu.zip

SHA-256 checksum793858e85694b65602841f52168bdbb2f9018f2a2e4e54003958fffd41ba729a

14972057 bytes

Linux ARM64 uses Vulkan for GPU acceleration; CUDA is not required. Jetson Orin compatibility and runtime verification are not yet qualified.

Release metadata and checksums

Extract. Open a terminal. Solve.

Install the standalone tool

  1. Download the archive matching your computer.
  2. Extract the archive into a folder you want to keep.
  3. Open Terminal on macOS or Linux, or PowerShell on Windows, in that folder.
  4. Run ./loom-solve --help on macOS or Linux, or .\loom-solve.exe --help in PowerShell.

The macOS package is built for Apple silicon with a macOS 14.0 deployment target.

On Windows, install the Microsoft Visual C++ v14 Redistributable (x64) before running the tool.

The Linux builds target glibc 2.35 using an Ubuntu 22.04 build baseline. The ARM64 package targets Jetson Orin through Vulkan; physical-device qualification is pending.

The standalone tool does not use the PixInsight update repository. For the native PixInsight process, follow the PixInsight installation instructions.

Check your download

Compare the archive’s SHA-256 with the checksum next to your download before extracting it.

macOS
shasum -a 256 PATH_TO_ARCHIVE
Linux
sha256sum PATH_TO_ARCHIVE
Windows PowerShell
Get-FileHash PATH_TO_ARCHIVE -Algorithm SHA256

Replace PATH_TO_ARCHIVE with the downloaded file’s path.

Your first solve.

Use a two-dimensional monochrome or three-plane RGB FITS image (primary HDU). No starting coordinate or local star catalogue is required.

macOS and Linux

Print the exact WCS as JSON:

./loom-solve image.fits --json

Save the result next to your image:

./loom-solve image.fits --json --sidecar

Windows PowerShell

Print the exact WCS as JSON:

.\loom-solve.exe image.fits --json

Save the result next to your image:

.\loom-solve.exe image.fits --json --sidecar

Replace image.fits with your image’s path. Sidecar output writes a new image.fits.wcs.json file. Existing files are never overwritten. Results go to standard output; progress and errors go to standard error. JSON preserves the exact WCS, including spline control points; spline solutions require JSON in this preview.

GPU acceleration is enabled by default, with CPU fallback available. Add --no-gpu for CPU-only execution. Optional --gaia PATH files can accelerate catalogue access. First-time online catalogue access may need a network connection; --offline restricts access to embedded, local and cached data.

Command reference.

The command’s own help, published with the release.

Loom PlateSolve — blind astrometric solver

Usage: loom-solve [OPTIONS] <IMAGE.fits>

Input:
  -i, --input PATH       Input FITS (alternative to positional path)
      --gaia PATH        Local Gaia XPSD file; repeatable (optional accelerator)
      --offline          Use embedded/local/cached catalogues only
      --ra DEG --dec DEG  Approximate centre (ICRS degrees; supply both)
      --scale ARCSEC     Approximate image scale in arcseconds/pixel
      --threads N        Worker count (default: automatic)

Output:
      --json             Print exact WCS as JSON instead of FITS WCS cards
  -o, --output PATH      Write result to a new file instead of stdout
      --sidecar          Write IMAGE.fits.wcs (or .wcs.json with --json)
      --write-wcs        Write solved FITS to --output (requires new output path)

Execution:
      --gpu              Enable GPU acceleration (default; CPU fallback available)
      --no-gpu           Use CPU only
  -q, --quiet            Suppress progress logs; errors remain on stderr
  -v, --verbose          Include detailed solver diagnostics
      --log-level LEVEL  off, error, warn, info, debug, trace
      --color MODE       auto, always, never (default: auto; honors NO_COLOR)
  -h, --help             Show this help
  -V, --version          Show version and release source revision

Results use stdout; logs and errors use stderr. Existing files are never overwritten.
LOOM_SOLVE_GAIA supplies OS-separated paths when --gaia is omitted.
Exit codes: 0 success, 1 unsolved, 2 usage, 3 input, 4 output, 130 cancelled.
Download the help reference