Skip to main content

Download

Get the engine. It's one line.

The installer downloads the prebuilt openeng binary, verifies its SHA-256 checksum, and puts it on your PATH. A default open-weight model is pulled on first run; each config then downloads whatever models it references from its registry — Hugging Face, the Ollama library, ModelScope, or a direct URL.

Linux x86_64 (and WSL) — live today

curl -fsSL https://openeng.ai/engine/install.sh | bash

One binary

A single Rust binary: the agent loop, the model runtimes, and the memory system — all in one file.

Open-weight models

Downloads open-weight models from any registry and runs them isolated in a sandbox on your machine. After the download, inference is offline.

Verified

TLS download, published SHA-256, and a script that refuses anything that doesn't match.

Prefer to read before you pipe to bash? Inspect install.sh.

Platform status

engine binary availability per platform
PlatformBinaryStatus
Linux · x86_64openeng-linux-x86_64Live
macOS · Apple Siliconopeneng-darwin-aarch64Coming soon
Windows · x64openeng-windows-x86_64.exeComing soon

No sign-up-to-be-notified forms — when a build ships it appears here and on releases.openeng.ai. Windows users can run the Linux binary under WSL today.

Manual install

Download openeng-linux-x86_64, verify it with sha256sum, mark it executable, and move it onto your PATH.

SHA-256: /engine/bin/openeng-linux-x86_64.sha256

After the install

Sign in once, then it's local

  1. 1Run openeng. Choose “Login through browser” — with a device-code fallback for headless machines — or sign in with an API key.
  2. 2If the model download was skipped or interrupted, the engine finishes it on first launch.
  3. 3Your Console-published config resolves automatically — agents, models, cells, MCP servers.
  4. 4Type a prompt, or /help for the command reference. Free covers the full on-device agent.

Full walkthrough in the docs — models and runtimes, threads, and permissions.

Also coming

The OpenEng desktop app

A desktop companion to the engine is being packaged for macOS, Windows, and Linux. No installers are published yet — these cards go live the moment a build ships, and releases land on the releases page first.

Coming soon

macOS

macOS 11+ · Apple Silicon · DMG

Coming soon — the macOS build is not published yet. Use the engine today.

Coming soon

Windows

Windows 10 / 11 · x64 · EXE

Coming soon — the Windows build is not published yet. Use the engine today.

Coming soon

Linux

Most distros · x86_64 · AppImage

Coming soon — the Linux desktop build is not published yet. The Linux engine is live.

Release announcements land on /releases, fed by releases.openeng.ai .

The engine is the product. It's ready.

Don't wait for an installer with a progress bar — the agent ships as one verified binary, today.

Linux x86_64 · checksum-verified

curl -fsSL https://openeng.ai/engine/install.sh | bash

Installing on servers? See headless mode and API keys.