Master the interactive CLI with keyboard shortcuts, slash commands, and file mentions
Interactive mode is the primary way to work with Cline CLI when you want a collaborative, conversational experience. Unlike headless mode (which runs a single task and exits), interactive mode keeps a session open where you can have back-and-forth conversations with Cline, refine your requests, and guide the AI as it works.
Before using interactive mode, you need to have Cline CLI installed and authenticated. If you haven’t done this yet, follow the Installation & Setup guide first.
Keyboard shortcuts are the primary way to navigate and control the interactive CLI. Since there’s no mouse interaction in the terminal, learning these shortcuts will help you work efficiently and switch between modes, manage input, and control your session without breaking your flow.
If you have workflows configured, they appear as additional slash commands. For example, if you have a workflow named code-review, you can invoke it with:
Cline operates in two modes, toggled with Tab. These modes work the same way in the CLI as they do in the VS Code extension. For a deeper explanation of how Plan and Act modes work, see the Plan and Act documentation.
Press Shift+Tab to toggle auto-approve for all actions. This removes the approval prompts that appear before each action, letting Cline work continuously without interruption.
When enabled, these actions happen without prompting:
File reads
File writes
Command execution
Browser actions
MCP tool calls
You can also configure granular auto-approve settings (e.g., auto-approve reads but not writes) via /settings → Auto-approve tab, or see the Auto-approve documentation for more details.
Auto-approve gives Cline full autonomy. Use on a clean git branch so you can easily revert changes if needed. You can always press Ctrl+C to stop Cline immediately.
By default, all CLI instances share the same settings and state. However, you may want to run isolated instances with separate configurations for scenarios like:
Different models for different tasks - Use a fast, cheap model for quick questions in one terminal and a more capable model for complex refactoring in another
Separate work and personal projects - Keep API keys, rules, and task history isolated between contexts
Testing configuration changes - Experiment with new settings without affecting your main setup
Team vs. individual settings - Use shared team configuration for work projects and personal preferences for side projects
To run isolated instances, use the --config flag with different directories:
Copy
Ask AI
# Work instance with team configurationcline --config ~/.cline-work# Personal instance with different model/providercline --config ~/.cline-personal# Experimental instance for testing new settingscline --config ~/.cline-test
Each config directory maintains its own provider settings, API keys, task history, and preferences.
Use terminal multiplexers like tmux or split terminals to run multiple Cline instances in parallel, each working on different parts of your project with different models or settings.
The interactive CLI maintains conversation context. Build on previous messages:
Copy
Ask AI
> Add a login endpoint[Cline creates the endpoint]> Now add rate limiting to it[Cline modifies the same endpoint]> Add tests for both features[Cline creates test files]
