Connecting your coding agents
AgentDock detects agent CLIs you've already installed and authenticated — it doesn't install or authenticate them for you.
AgentDock connects to agent CLIs that are already installed and authenticated on your machine. It does not bundle, install, or manage credentials for any of them — each agent is a separate product with its own installation and sign-in process, entirely outside AgentDock's control.
What AgentDock actually does
At launch (and when you open Settings → Agents), AgentDock searches your PATH — plus a few known install locations on Windows and macOS — for these executables:
| Agent | Executable | Notes |
|---|---|---|
| Claude Code | claude | Detected via a standard PATH search. |
| OpenAI Codex | codex | Uses a dedicated multi-step resolver, not a plain PATH search — see the Codex page for why. |
| Google Antigravity | agy (also probes antigravity, google-antigravity) | agy is the confirmed real executable name. |
Detected agents appear as options when you start a session. Settings → Agents also lets you override the executable path per agent and run a live "Test" to confirm AgentDock can launch it.
What you need to do yourself
- Install each agent CLI following that provider's own instructions.
- Authenticate it — run its sign-in flow outside AgentDock, the same way you would from any terminal.
- Confirm it's on your
PATH(or set a custom path in Settings → Agents).
Note
Model access, usage limits, and billing are between you and each agent's provider (Anthropic, OpenAI, or Google) — AgentDock has no visibility into or control over that relationship.
If AgentDock doesn't detect a CLI you know is installed, see Agent CLI not detected. For how AgentDock verifies you're signed in, see Authentication.