Get Orrery

Private beta

Download the preview for Windows (x64), Linux (x64 or ARM64) or macOS (Apple silicon). Sign in (free) to download. Local models run free; hosted models need Pro.

Free with your own modelsRun the agent on Ollama, LM Studio, vLLM or llama.cpp on your computer or local network. No subscription, no API key.

Pro $40 / Max $100 / Ultra $200 /moRun governed agents with open-source models on your device plus hosted-model credits and Orrery Cloud.

Paid tiers auto-renew monthly via Stripe; cancel anytime. Hosted-model prompts are proxied through Orrery Cloud. Terms | Privacy.

Request access

Pro is $40/month, cancel anytime. Anyone with a free account can download and run agents on their own models; a paid plan adds hosted models. For beta questions, email kt@ephemerent.com.

Download the preview for Windows (x64), Linux (x64 or ARM64) or macOS (Apple silicon). Sign in (free) to download. It runs your local models for free; Pro adds hosted models, any provider with your own key, and Orrery Cloud. Preview software may change, fail, or produce incorrect output.

Download Orrery

Other platforms

Linux: the .deb installs cleanly on Ubuntu and Debian. The AppImage needs libfuse2, which Ubuntu 24.04 no longer ships (sudo apt install libfuse2t64), or run it once with --appimage-extract-and-run. Windows: unzip and run Orrery Test Beta.exe; SmartScreen will warn because the preview is unsigned.

Windows local speech recognition and neural voices also need the Microsoft Visual C++ v14 Redistributable (x64).

Downloads need a free account, and each link expires ten minutes after you request it. Preview builds are unsigned: on macOS move Orrery to Applications, open it once, then choose Open Anyway in System Settings, Privacy & Security (if macOS calls it damaged, run xattr -dr com.apple.quarantine /Applications/Orrery.app in Terminal); on Windows choose More info → Run anyway; on Linux chmod +x the AppImage.

Release candidatePending verification

Update modeWebsite-hosted updates

Verify downloadSHA-256 on release page

Release notes and checksums

Free with your own modelsRun the agent on Ollama, LM Studio, vLLM or llama.cpp on your computer or local network. No subscription, no API key.

Pro: $40/moHosted models (DeepSeek, GLM and others) with credits included, any provider with your own key (OpenRouter, OpenAI, Anthropic…), and Orrery Cloud. Max $100 and Ultra $200 add bigger pools.

Paid tiers auto-renew monthly via Stripe; cancel anytime. Credits reset monthly and do not roll over. Hosted-model prompts are proxied through Orrery Cloud. Agent output is not guaranteed; keep backups and review changes before use. Terms | Privacy.

Quick start

Download for your platform and open it. The preview is not code-signed yet: on Windows choose More info, then Run anyway; compare the checksum if you want to be sure, and never turn off your system's protections. On macOS, open it once, then choose Open Anyway in System Settings, Privacy & Security.
Sign in with the same account you used here, or skip sign-in and pick a local model. A paid plan or complimentary access unlocks hosted models; they need no API keys.
Open a project folder (or the built-in sample), describe a task in plain words, and watch it plan, edit, and run your tests.
Read the diff and the receipt before you keep anything: what changed, what passed, and what it cost.
Linux installs update themselves. On Windows and macOS, close the app and download the newest zip from this page when a release notes a fix you want.

Need help or want to share feedback? Join the Discord.

Identify (optional)

Email only — stored on this device. Skip if you don't need named entries in your local activity log.