Orrery

Orrery

[繁體中文](docs/README-zh_TW.md)

**Orrery is a runtime environment manager for AI tools.**

It lets you run Claude Code, Codex CLI, and Gemini CLI across isolated environments — each with its own account and credentials — while keeping your conversations continuous across account switches.

The CLI command is lowercase `orrery`. The product name is capitalized **Orrery**.

---

🧠 Why Orrery?

Working with AI CLI tools today is messy:

• Switching accounts breaks your context
• Conversation history doesn't follow you across accounts
• Tools can't coordinate tasks between environments

Orrery solves this by introducing:

**Isolated, composable AI environments**

Each environment has its own auth credentials and configuration. But sessions — your conversation history and project context — are **shared by default**, so you can switch accounts and pick up exactly where you left off.

---

🧩 Core Concepts

Start with **accounts** — that's all most people need. Reach for a **workspace** only when a context needs fully isolated config.

Account

Your **identity** for a tool — the credential Orrery logs in with. Accounts live in a shared pool: add once, switch any time with `orrery use`. This is the layer you'll touch every day.

Workspace _(advanced)_

An optional **isolation** layer: separate memory, sessions, and env vars on top of accounts. Most users never need one — reach for a workspace when a client or project needs its own config space. Pin an account to it with `orrery pin --workspace `, then activate it with `orrery use `.

Session

Represents **continuity**: conversation history and project context. Shared across account switches by default — `claude --resume` just works after switching accounts.

Phantom mode

Running `claude` launches it under a phantom supervisor. From inside that Claude session, t