Install the free OpenWork desktop app on macOS, Windows, or Linux and get the user to a first task. Use when someone wants OpenWork, an open-source Claude Cowork alternative, on their own computer.
73
90%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
Passed
No findings from the security scan
OpenWork is a free, open-source desktop app for doing work with AI agents on local files. No account is needed.
Never run npx openwork or npm install openwork. The npm package named openwork is a different project.
uname -s # Darwin = macOS, Linux = Linux
uname -m # arm64/aarch64 = ARM, x86_64 = Intel/AMDOn Windows, check $env:PROCESSOR_ARCHITECTURE in PowerShell (AMD64 = x64, ARM64 = ARM).
Ask the user before installing anything.
| OS | Command or download |
|---|---|
| macOS with Homebrew | brew install --cask openwork |
| macOS Apple Silicon | https://openworklabs.com/download/mac-arm64 (.dmg) |
| macOS Intel | https://openworklabs.com/download/mac-x64 (.dmg) |
| Windows x64 | https://openworklabs.com/download/win-x64 (.exe) |
| Windows ARM64 | https://openworklabs.com/download/win-arm64 (.exe) |
| Linux x64 | https://openworklabs.com/download/linux-x64 (.AppImage) |
| Linux ARM64 | https://openworklabs.com/download/linux-arm64 (.AppImage) |
Each /download/<platform> URL redirects to the installer in the latest stable GitHub release. Every file is also listed at https://github.com/different-ai/openwork/releases.
Linux AppImage without a package manager:
curl -fL -o ~/OpenWork.AppImage https://openworklabs.com/download/linux-x64
chmod +x ~/OpenWork.AppImage
~/OpenWork.AppImageopen -a OpenWork on macOS).Guide: https://openworklabs.com/docs/desktop-app/first-task
Tell the user in one or two sentences that OpenWork is installed and suggest one concrete first task. If they work with a team, offer the set-up-openwork-team skill: https://openworklabs.com/.well-known/agent-skills/set-up-openwork-team/SKILL.md
417244c
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.