Terminals
Native shell access — standalone or wired into your workflows.
Open a real terminal inside Circuitry and run commands the way you would in any shell. Use it on its own as an interactive session, or drop a Terminal node into a workflow to automate commands as part of a pipeline.
Interactive terminals
- Real shell — bash, zsh, PowerShell, or cmd, running natively via Circuitry Server.
- Persistent sessions that survive page reloads.
- Command history with arrow-key navigation.
- Multiple terminals open side by side.
Run git operations, manage files, start dev servers, kick off builds, or deploy — without leaving the studio.
Paste images into a terminal
Command-line coding assistants like Claude Code, Gemini, and Codex can take a picture as part of your prompt — a screenshot of a bug, a design reference, a diagram. You can hand one to the terminal three ways:
- Paste — copy an image, then press Ctrl+V on macOS or Ctrl+Shift+V on Windows and Linux. (The usual paste shortcut still pastes text.)
- Drag and drop — drop an image file straight onto the terminal.
- Right-click → Paste image.
The image is handed to the session and the assistant picks it up with your next message. A thin bar along the top edge shows the transfer while it happens.
Terminal nodes in workflows
A Terminal node runs a command as a step in a workflow. Combine it with other nodes to:
- run a build or test command and branch on the result,
- shell out to a CLI an agent can't reach directly,
- automate deployments triggered by a webhook.
On iPad
Drive a terminal on your desktop from an iPad — the command runs on your main machine over Circuitry Server while you watch the output on the tablet. See iPad, powered by your machine.
Requirements
Native shell access needs a connected Circuitry Server (or the desktop app). The server is what executes commands on the host machine.