Cline

Cline is an AI assistant that works alongside your command line and code editor. It uses agent-style programming powered by large language models to break complex development tasks into steps and execute them in a controlled way. Cline supports the following capabilities:

• Create and edit files, and browse large projects with model-driven actions.
• Use the browser and run terminal commands, with your confirmation before each action.
• Extend its capabilities through the Model Context Protocol (MCP), including custom tools.
• Unlike sandboxed scripts, Cline provides a visual interface and asks for confirmation before each file change or terminal command, helping keep AI-assisted development safe and controlled.

Common Cline use cases include:

• Completing multi-step refactoring or feature implementation tasks.
• Exploring and modifying codebases with natural-language instructions.
• Automating repetitive development workflows, such as generating tests or updating configuration files, while keeping human approval in the loop.

​

Configure

The steps below use Cline 3.20.1 in VS Code as a reference. Newer versions may differ slightly.

​

1. Install the extension

Open VS Code, or another supported editor, and install the Cline extension from the marketplace.

​

2. Configure the model

In Cline settings, fill in the following fields. Adjust other model parameters as needed. For complex code-editing or long-running tasks, choose a model with sufficient context length. If applicable, also check whether the model supports reasoning.

​

Verify

After saving the configuration, submit a task to Cline in natural language. Cline calls the ScitiX Model Inference API and executes file or terminal actions after your confirmation.

​

FAQ

• Cannot use checkpoints in Desktop directory Cline does not allow checkpoints when the workspace is under the Desktop folder. Open your project from another directory, such as ~/Code, ~/Projects, or another development folder, to avoid this error.