Getting Started

Installation

Install the dxflow engine on Linux, macOS, or Windows

dxflow ships as a single binary for Linux, macOS, and Windows on x86_64 and ARM64. Install it with the script, or download the archive and place the binary on your PATH.

System requirements

dxflow is lightweight and runs on minimal resources, making it suitable for everything from IoT devices to high-performance clusters.
RequirementMinimum
OSLinux, macOS, or Windows
Architecturex86_64 (AMD64) or ARM64
Memory512MB RAM
Storage100MB disk space
NetworkInternet connection for installation
RuntimeDocker, Podman, Singularity, or Apptainer — needed to run workflows

A runtime is needed only to run workflows. The engine itself, its CLI, its API, shells, and artifacts all work without one.

Install with the script

The fastest route on Linux and macOS. The script detects your OS and architecture, downloads the matching release, installs it to /usr/local/bin, and sets the execute bit.

The script installs to /usr/local/bin and requires root. Review the install script before running it.
curl -fsSL https://dxflow.ai/install.sh | sudo bash

Jump to verify installation once it finishes.

Install manually

Download

Pick your platform and architecture from the GitHub releases page, or use a direct link:

x86_64

Linux

macOS

Windows

ARM64

Linux

macOS

Windows

Every release ships a dxflow_checksums.txt. Verify your download against it before extracting:

wget https://github.com/dxflow-ai/community/releases/latest/download/dxflow_checksums.txt

sha256sum --check --ignore-missing dxflow_checksums.txt      # Linux
shasum -a 256 --check --ignore-missing dxflow_checksums.txt  # macOS

Extract

Every archive is a .tar.gz holding one dxflow binary (dxflow.exe on Windows).

Extract to a directory on your PATH

sudo tar -xzf dxflow_linux_amd64.tar.gz -C /usr/local/bin

Grant execute permission

sudo chmod +x /usr/local/bin/dxflow

Verify installation

Open a new terminal and check the binary:

dxflow --version
dxflow --help

Then confirm the engine runs on this machine and see what it detected:

dxflow engine info

This command spawns a temporary engine when none is running, so it works right after install:

● Engine
  Identity     a1b2c3d4e5f60718293a4b5c6d7e8f90
  Directory    ~/.dxflow
  Profile      default
  Version      2.4.0
● System
  OS           Ubuntu 24.04  (amd64)
  CPU          AMD EPYC 7763  (16 CORE)
  GPU          none
  Memory       64 GB
  Runtime      docker 27.3.1
The Runtime row lists the container runtimes dxflow found. none means workflows have nothing to run on yet — install Docker, Podman, Singularity, or Apptainer.

Next steps

Quick Start

Boot the engine and run your first workflow

CLI Reference

Every command, flag, and option

Troubleshooting

ProblemFix
dxflow: command not foundConfirm the install directory is on your PATH, then open a new terminal.
permission denied running dxflowsudo chmod +x /usr/local/bin/dxflow
macOS blocks the binaryApprove it in System SettingsPrivacy & SecurityOpen Anyway.
Download failsCheck connectivity and that the releases page is reachable, then download and extract manually.
Checksum mismatchDownload again — the archive is incomplete or corrupted.

For more, see the FAQ and Troubleshooting.