Overview

You can use our AI platform in agentic mode in VSCode using the Cline plugin.

Cline is an open-source AI coding assistant for Visual Studio Code that acts as an autonomous software engineering agent. Unlike traditional code completion tools, Cline can understand your project, edit multiple files, execute terminal commands (with your approval), and help implement complex development tasks.

The RM2Lab AI Platform provides an OpenAI-compatible endpoint that can be used directly from Cline, allowing you to leverage the locally hosted Qwen3.6-27B coding model for everyday software development.

Prerequisites

Before configuring Cline, you will need:

• A VS Code installation
• The Cline extension installed from the Visual Studio Code Marketplace
• An RM2Lab AI Platform account
• A personal API key generated from Open WebUI

Official Cline documentation:

• https://docs.cline.bot/
• https://docs.cline.bot/getting-started/installing-cline 

Getting an API Key

API keys are managed through Open WebUI.

1. Open:

   
   

   https://agent.rmlab.infn.it

2. Login using INFN AAI.
3. Click your profile icon in the top-right corner.
4. Open Settings.
5. Navigate to API Keys.
6. Generate a new personal API key.

Keep the generated key secure, as it provides access to your account.

Configuring Cline

Open the Cline settings and select an OpenAI-compatible provider.

Use the following configuration:

The model name is case-sensitive and must be entered exactly as:

code

Available Model

Cline is configured to use the following coding model:

The model has been optimized for:

• Software development
• Code generation
• Refactoring
• Debugging
• Documentation
• Code review
• Architecture discussions
• General programming assistance

The configured context window is 150k tokens, allowing the model to reason over very large codebases or multiple source files within a single conversation.

What Cline Can Do

Once configured, Cline can assist with a wide variety of development tasks, including:

• Understanding large projects
• Explaining existing code
• Implementing new features
• Refactoring existing code
• Writing unit tests
• Generating documentation
• Reviewing pull requests
• Detecting bugs
• Suggesting architectural improvements
• Assisting with DevOps and automation scripts

Unlike traditional code completion assistants, Cline can work across multiple files, inspect your project structure, interact with the terminal (upon approval), and execute multi-step development workflows. 

Recommended Usage

For the best experience:

• Keep your project opened as a VS Code workspace.
• Let Cline inspect the existing codebase before requesting large changes.
• Break very large tasks into smaller, incremental requests.
• Review and approve generated file modifications before applying them.
• Use Git to track and review AI-generated changes.

Notes

• API keys are personal and must not be shared.
• Usage is subject to the RM2Lab AI Platform availability.
• The platform is currently experimental and may undergo updates, model changes, or temporary service interruptions without prior notice.
• The code model is continuously updated as newer and more capable coding models become available on the RM2Lab infrastructure.