Labelbox unveils integrated VS Code IDE: Generate sophisticated training code quickly

We are thrilled to announce a significant enhancement to the Labelbox Platform: the integration of a full Visual Studio Code (VS Code) Web IDE environment directly within our multimodal chat editor. The Labelbox Platform now delivers a desktop-class coding experience, enabling AI labs to accelerate the development of frontier models thanks to higher-quality code training data.

By combining a powerful IDE with our global Alignerr network of expert AI trainers and coders, we can now produce sophisticated coding data faster and with greater precision—unlocking new levels of efficiency and quality. The robust, familiar coding environment helps developers and AI trainers tackle the complex coding projects with ease. 

The crucial role of code in frontier AI training

Coding has emerged as a cornerstone use case for training frontier AI models, particularly LLMs. As generative AI becomes a vital tool for software developers—assisting with code generation, debugging, and optimization—models must be trained on advanced, real-world coding scenarios to deliver meaningful results. This requires more than basic snippets; it demands exposure to complex codebases, multi-file projects, and the tools developers rely on daily.

The challenge lies in creating training data that mirrors the environments where developers work. Lightweight editors fall short when handling large files or intricate repositories, limiting the ability to generate high-fidelity datasets. To meet the needs of AI labs, efficient tools that replicate real-world IDEs are essential. By providing Alignerrs with a robust coding environment, Labelbox ensures that models can learn from the kind of sophisticated, context-rich data that drives innovation in software development.

Introducing the integrated VS Code web IDE environment

Our latest release integrates a full Visual Studio Code (VS Code) Web IDE into the Labelbox Platform multimodal chat editor (MMC), transforming how users interact with code. Previously, MMC offered the lightweight Monaco editor for code labeling—a solid foundation, but not a substitute for a rich, desktop-class IDE. 

Now, with the click of the “Edit in VS Code” button, users can launch a full IDE workspace: a remote host environment powered by VS Code. This upgrade allows seamless work on entire code repositories, execution of CLI tools, debugging, test writing, and installation of additional VS Code extensions—all enhanced by the seamless availability of Github Copilot for AI-driven code assistance.

When users activate this feature, the system spins up a fresh workspace and automatically imports any existing code being reviewed or written, creating multiple files as needed. This means you can edit large files, manage multi-file projects, and leverage familiar VS Code capabilities without leaving the Labelbox ecosystem. The result is a streamlined workflow that saves time and boosts productivity for coding tasks.

Key highlights include: 

• A full VS Code Web IDE instance: Access the complete suite of VS Code features directly within the Labelbox Platform, including debugging, terminal access, and extension support.
• Remote host environment: Run CLI tools, execute code, and manage dependencies in a dedicated environment.
• Seamless integration: Open the IDE with a single click via the "Edit in Coder" button in MMC, automatically loading existing code from your project.
• AI-powered assistance: Leverage Github Copilot for AI-driven code completion and generation, enhancing productivity and code quality.
• Support for entire code repositories: Work across multiple files and complex projects with ease.

The Alignerr advantage: Expert coders, enhanced workflows

A key differentiator of Labelbox is the combination of our powerful Labelbox Platform with our Alignerr network, a global community of highly skilled AI trainers spanning multiple languages and domains including STEM, finance, legal, speech, medical, and more. This network is pivotal in helping AI labs operate or staff their modern AI data factory, leading to the creation of original and customized training data. 

With the new VS Code IDE, they can develop and review code at an even higher standard. The Labelbox Platform’s customizable review workflows, paired with Alignerr’s talent, ensure rigorous quality assurance (QA) for every project. This fully-managed service sets us apart, delivering not just tools but a comprehensive solution for AI labs seeking top-tier training data.

The integrated IDE will also play a role beyond project work. In the future, we plan to utilize it for Alignerr coding assessments, enabling us to evaluate and onboard trainers with the precise skill levels needed for your initiatives. This will further ensure that your data factory is staffed with the right expertise, every time.

How to get started

Getting started with the new feature is straightforward. To access the VS Code Web IDE, create a project in Labelbox using the multimodal chat editor. When editing a code block, simply click the “Edit in Coder” button. This launches a fully-functional VS Code instance in a remote workspace, pre-populated with any code already present in your label. From there, you can:

• Run and debug code using standard commands (e.g., Python scripts with terminal inputs)
• Install VS Code extensions to tailor the environment to your needs
• Use Github Copilot—authenticated via GitHub credentials—for AI-assisted coding
• Manage multi-file projects and execute CLI tools seamlessly

Once your work is complete, copy the updated code back into your MMC project with ease. 

Empowering the future of AI coding

The integration of a VS Code Web IDE into the Labelbox Platform marks a significant leap forward for AI labs building frontier models. By combining a best-in-class coding environment with the expertise of our Alignerr network, we’re enabling teams to produce higher-quality training data faster than ever before. 

Ready to revolutionize your AI training?

• Sign up for a Labelbox account today to experience the integrated VS Code web IDE firsthand.
• Contact our sales team to discuss how we can help operate or staff your data factory with our Alignerr network.