Build Multiple Projects in Parallel with Verdent AI 🔥

This video dives into Verdent, an AI coding environment that lets you plan, code, and verify projects with confidence. It highlights Verdent’s unique features like parallel agent execution and work trees, setting it apart from other AI coding tools.

Key Takeaways

• Parallel Project Development: Verdent allows you to work on multiple projects simultaneously, a feature not commonly found in other AI coding environments.
• Work Tree Feature: This game-changing feature enables isolated development environments within projects, allowing for safe experimentation and easy merging.
• Granular Control: Verdent offers extensive control over AI models, user rules, and deny rules, giving developers fine-tuned command over the AI’s behavior.
• Speed and Efficiency: The platform is noted for its impressive speed in generating code and plans, significantly speeding up the development workflow.
• VS Code Integration: Verdent seamlessly integrates with VS Code through an extension, complementing the existing IDE without replacing it.

Getting Started with Verdent

Verdent offers both a desktop app and a VS Code extension. The desktop app runs at the operating system level, while the extension works within your VS Code IDE. To get started, you can download the installer from their website. Once installed, you can log in to your Verdent account.

The VS Code extension appears as a new tab within your IDE. It offers three modes: Agent, Plan, and Chat. You can also configure how the AI operates, choosing between manual acceptance, automatic running, or having the AI handle all operations. This flexibility allows you to tailor the AI’s interaction to your workflow.

Planning and Code Generation

Verdent’s plan mode is quite in-depth. When you ask it to create something, like a Next.js app for an invoice generator, it asks clarifying questions about core features. This interactive approach mimics how a human engineer would gather requirements.

For example, when building the invoice generator, the AI asked about desired features such as adding/editing invoice items, client company information, text discounts, PDF export, and preferred technologies like Tailwind CSS and JavaScript. After receiving these details, Verdent generates a detailed plan, including a graph of the application’s structure, step-by-step tasks, and state definitions. You can then click "Start Building" to have the AI generate the code based on this plan.

Unique Features: Work Trees and Parallelism

One of Verdent’s standout features is its work tree concept. This allows you to create isolated branches or workspaces within a project. For instance, if you want to experiment with making an app more colorful without affecting the main codebase, you can create a new "feature" workspace. This is similar to creating a new branch in Git, but it’s a fully isolated environment within Verdent. You can test your changes, and if you like them, merge them back into your main project. If not, you can simply discard the workspace without any impact on your base project.

This leads to Verdent’s true parallelism. The desktop app truly shines here, allowing you to run multiple projects simultaneously. You can initiate the creation of several different applications at once – for example, a PDF utility app, a sticky notes app, and a basic text editor. The AI will work on all of them concurrently. This level of multitasking is impressive and can significantly speed up development, especially when you have multiple ideas to pursue.

Practical Applications and Demonstrations

The video showcases several practical examples:

• Invoice Generator: Verdent successfully planned and began generating code for a Next.js invoice generator based on detailed user requirements.
• Sticky Notes App: The AI created a functional sticky notes app with three columns (Pending, Doing, Done) and drag-and-drop functionality. It also demonstrated persistence using local storage after a refresh.
• PDF Utility App: Verdent built a PDF utility app capable of merging, rotating, splitting, and editing PDF metadata. The demonstration showed its ability to split a PDF by specific page ranges, proving its effectiveness.
• Basic Text Editor: A simple text editor app was created, allowing users to write and save .txt files, with data persisting in local storage.

Granular Controls and Settings

Verdent’s settings panel offers a high degree of control. You can switch between light, dark, or system themes. It supports proxy settings and allows you to select specific AI models for different tasks like general agent work, code review, and plan mode. This granular control extends to per-model settings, where you can define user rules (e.g., preferring a purple theme) and deny rules to prevent certain actions.

Furthermore, Verdent includes built-in sub-agents like a Verifier and a Plan Reviewer, and you can create your own custom sub-agents. This modular approach enhances the AI’s capabilities and allows for specialized workflows.

Conclusion

Verdent stands out as a powerful AI coding agent, particularly for its true parallelism and work tree features. It enables developers to manage multiple projects and tasks simultaneously, experiment safely in isolated workspaces, and benefit from impressive speed and granular control over AI models. While competitors like VS Code with GitHub Copilot, Antigravity, and Warp exist, Verdent’s unique approach to multitasking and isolated development makes it a compelling choice for developers looking to boost their productivity.