Shepherd Terminal

What is Shepherd Terminal?

Shepherd Terminal is a terminal workspace built around the way developers now run multiple AI coding agents. It lets users split agents across tabs and panes, keep terminal processes alive after closing the app, and reconnect to the same live sessions later. The product includes an agent monitor that shows which sessions are working, waiting, or completed, plus tools for inspecting files, diffs, and remote development routes. It is aimed at developers who use Codex, Claude Code, and remote machines as part of everyday engineering work.

How to use Shepherd Terminal?

1. Download and install the macOS app from the Shepherd website.
2. Create tabs or panes for the coding agents and terminal sessions you want to run.
3. Start Codex, Claude Code, or another CLI agent inside the workspace.
4. Use the Agent Monitor to see which sessions are working, waiting, or finished.
5. Reconnect later to resume the same live sessions and inspect files or diffs.

Core Features

• Persistent terminal sessions - Keeps agent and shell processes alive when the app closes or reconnects.
• Agent-focused panes - Runs multiple coding agents across tabs, panes, and machines.
• Live Agent Monitor - Shows active, waiting, completed, and unread sessions in one place.
• Workspace-aware plugins - Lets agents create panes, read output, send input, and arrange their own workspace.
• Remote development support - Connects to remote machines while forwarding routes and browsing files locally.
• Change inspection - Helps users return to an agent session and review the files and diffs it created.

Use Cases

• Parallel agent development - Run several Codex or Claude Code tasks without losing track of status.
• Long-running coding work - Keep terminal jobs alive across app restarts or dropped connections.
• Remote machine workflows - Start agent work on a remote server and monitor it from a local Mac.
• Agent review loops - Inspect files, diffs, and browser feedback alongside the terminal session.