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
92%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
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
different-ai/openwork
acacc7b
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.