Standalone downloads
FX releases starting with 0.12.0 offer a native grida-fx executable for pinned CI images
and other tools that install the CLI directly. Each archive contains the executable and its
license. It includes the FX Viewer.
Choose your archive
Open the version you want on FX GitHub Releases.
Download its SHA256SUMS and the grida-fx-<version>-<target>.tar.gz asset for your machine.
Check the machine architecture with uname -m.
| Platform | Architecture | Target |
|---|---|---|
| macOS 11 or later | Apple silicon (arm64) | aarch64-apple-darwin |
| macOS 10.12 or later | Intel (x86_64) | x86_64-apple-darwin |
| Linux with glibc 2.28 or later | aarch64 | aarch64-unknown-linux-gnu |
| Linux with glibc 2.28 or later | x86_64 | x86_64-unknown-linux-gnu |
On Windows, use the appropriate Linux archive in WSL 2. On Alpine Linux, use a glibc-based environment for the Linux executable.
Verify and install
From the download directory, set archive to the filename you downloaded. This example
uses 0.12.0 for Apple silicon.
archive=grida-fx-0.12.0-aarch64-apple-darwin.tar.gz
awk -v name="$archive" '$2 == name { print; count++ } END { exit count != 1 }' \
SHA256SUMS > selected.sha256 &&
shasum -a 256 -c selected.sha256
On Linux, you can use sha256sum -c selected.sha256 for the last command.
After the check reports OK, unpack into a new directory and install the executable.
mkdir fx-download
tar -xzf "$archive" -C fx-download
mkdir -p "$HOME/.local/bin"
install -m 755 fx-download/grida-fx "$HOME/.local/bin/grida-fx"
"$HOME/.local/bin/grida-fx" --version
Use the full path in CI, or add $HOME/.local/bin to your PATH. The guide's
grida fx commands take the same arguments when run as grida-fx.
Workflow hosts
The executable runs the CLI and serves the bundled viewer directly. Workflows with
Python nodes or built-in local processing need Python 3.11 or later with a matching
grida package. Workflows with TypeScript nodes need Node 24 or later and a matching
@grida/fx dependency. Getting started
explains how to select the Python or Node interpreter.