Terminal
PTY-backed terminal with clickable file paths, input history, and secret redaction.
Terminal Overview
Each agent execution gets a dedicated PTY-backed terminal. The terminal displays real-time output from the AI coding CLI, including tool calls, file edits, and command output. You can interact with the terminal directly to provide input when the agent pauses for confirmation.
Insert vs Submit Modes
Insert Mode
Type text into the input field. Press Enter to add a newline. Use the submit button or shortcut to send. Best for multi-line input and careful composition.
Submit Mode
Press Enter to send immediately. Best for quick responses like "yes", "no", or short commands.
Clickable File Paths
File paths in terminal output are automatically detected and rendered as clickable links. Click a file path to jump directly to that location in the file explorer, making it easy to navigate to files the agent mentions or modifies.
Input History
A session-backed input history panel stores your previous CLI inputs. Access it to quickly re-send commands or review what you have typed during the session.
Find in Terminal
Press Cmd+F to search terminal output. Useful for finding specific error messages, file paths, or output from long-running agent sessions.
Secret Redaction
API keys, tokens, and credentials are automatically redacted from terminal logs and stored output. This prevents sensitive values from being persisted in execution history or accidentally shared in screenshots.