Antigravity changes the way we build software. Instead of writing every line of code yourself, you tell smart AI agents what you want, and they handle the planning, coding, and problem-solving for you. It’s not just an autocomplete tool; it feels more like giving instructions from a Mission Control desk.

Your job also changes. You’re no longer stuck doing all the small tasks. You act more like a project lead who sets the direction, while the agents handle the detailed work. And if you ever want to jump in and edit the code yourself, you still can

In simple words, using Antigravity feels like moving from doing everything by hand to managing a smart team of digital workers. The diagram shows this clearly: the old way on the left and the new agent-driven way on the right.

Getting Started with Antigravity: A Guided Walkthrough

To help you make the most of this tutorial, we’ve organised the material into a simple, structured path:

1. Begin With Setup & Orientation

Start by installing Antigravity and applying the recommended configuration settings. This section also introduces the foundational concepts and essential navigation features you’ll rely on throughout your experience.

2. Explore Practical Use Cases

After completing the setup, browse through several example scenarios you can test right away—such as static and dynamic website creation, building interactive web applications, aggregating external news sources, and more.

By this point, you should have a clear sense of how Antigravity operates. It’s a perfect moment to try out your own prompts and tasks.

3. Customise Antigravity to Fit Your Workflow

Next, dive into personalising how Antigravity behaves. Here you’ll learn about Rules and Workflows, mechanisms that help you enforce coding standards, define reusable instructions, and trigger complex actions with a single command. Multiple examples are provided to guide you.

4. Configure Agent Security

Since Antigravity Agents may run a variety of shell or terminal commands, you may want to control what they can execute. This section walks you through setting Allowed and Restricted command lists, ensuring that sensitive actions require your approval while safe ones can run

Helpful Resources

Below is a quick reference to official Antigravity links

Installing Antigravity

Let’s start by installing Antigravity. The product is currently in preview, and you can begin using it with your personal Gmail account.

Visit the downloads page, then select and download the version that matches your operating system.

Setting up Antigravity

Launch the application installer and install the same on your machine. Once you have completed the installation, launch the Antigravity application. You should see a screen similar to the following:

Click on the next button. This brings up the option for you to import from your existing VS Code or Cursor settings. We will go with a fresh start.

The next screen is to choose a theme type. We will go with the Dark theme, but it's entirely up to you, depending on your preference.

The next screen is important. It demonstrates the flexibility that is available in Antigravity in terms of how you want the Agent to behave.

Let’s take a closer look at what these settings mean. Keep in mind that none of the choices you make here are permanent—you can adjust them at any point, even while the Agent is actively working.

Before exploring the available modes, it’s important to understand the two key settings displayed on the right side of the screen:

Terminal Execution Policy

This controls how freely the Agent can run terminal commands or tools on your machine. You can choose from three levels:

• Off:
  The Agent will never run terminal commands automatically, except those you explicitly place in an Allow List.
• Auto:
  The Agent evaluates each command and decides whether to run it automatically. When necessary, it will pause and ask for your approval.
• Turbo:
  The Agent automatically runs all terminal commands unless they are on a Deny List you configure.

Review Policy

While completing tasks, the Agent generates different items—such as task outlines, implementation plans, and more.
This policy controls who decides whether these items should be reviewed before the Agent continues.

You can choose from three behaviours:

• Always Proceed:
  The Agent never requests a review.
• Agent Decides:
  The Agent determines when a review is needed and prompts you accordingly.
• Request Review:
  The Agent always pauses and asks for your review before proceeding.

Now that these properties are clear, the four setup choices simply combine different execution and review behaviours into convenient presets. They help you define how much independence the Agent should have when running commands or advancing through its workflow.

Understanding the Four Antigravity Agent Modes

Antigravity gives you flexibility in how much control you want the Agent to have over your development workflow. Think of these modes as different “collaboration styles” between you and the Agent. Each one changes how independently the Agent works, when it asks for permission, and how much manual involvement you prefer.

1. Agent-driven development

Think of this as: “You tell the Agent what you want, and it runs with it.”

In this mode, the Agent takes the lead:

• It plans the task
• Executes terminal commands automatically
• Proceeds with minimal interruptions
• Only checks in when necessary

This is ideal when you want hands-off automation, similar to hiring a highly trusted assistant who handles end-to-end execution.

Best for:

• Fast prototyping
• Repetitive tasks
• Experienced users are comfortable with high autonomy.
  
  

2. Agent-assisted development (Recommended)

Think of this as: “A balanced partnership, you give direction, and the Agent keeps you in the loop.”

In this mode, the Agent works actively but keeps a healthy feedback loop:

• It proposes plans
• Executes most commands automatically
• Asks for your approval when needed
• Shares important artifacts with you

You stay informed without micro-managing. This mode offers the best mix of speed and oversight, which is why it’s recommended for most users.

Best for:

• Everyday development
• Users who want speed but still some visibility and control
• Teams that value accountability

3. Review-driven development

Think of this as: “You approve every step, like code review for every action.”

Here, the Agent moves deliberately:

• It always asks for your review before executing important steps
• You approve terminal commands, plans, and artifacts
• Nothing happens without your confirmation

This mode ensures maximum control and transparency, useful when accuracy or safety is critical.

Best for:

• Production-critical systems
• Highly regulated environments
• Developers who prefer detailed oversight

4. Custom configuration (Full manual control)

Think of this as: “Build your own workflow—fine-tune exactly how the Agent behaves.”

You can adjust:

• Terminal command permissions
• Review policies
• Safety list
• How autonomous or cautious the Agent should be

This mode is perfect if you have very specific preferences, want to experiment, or need a workflow that fits a unique development environment.

Best for:

• Power users
• Specialized workflows
• Teams with custom process guidelines
  

The Agent-assisted development mode offers the best balance of autonomy and oversight. It allows the Agent to make intelligent decisions while still checking in with you when approval is needed, which is why it’s the recommended option.

Go ahead and choose the mode that suits your workflow, though for now, the recommended setting is a great place to start.

Next, you’ll move on to setting up your editor. Select the themes and preferences that match your style.

As mentioned earlier, Antigravity is available in preview mode and free if you have a personal Gmail account. So sign in now with your account. This will open up the browser, allowing you to sign in.

On successful authentication, you will see a message similar to the one below, and it will lead you back to the Antigravity application. Go with the flow.

The last step, as is typical, is the terms of use. You can make a decision if you’d like to opt in or not, and then click on Next.

This will lead you to the moment of truth, where Antigravity will be waiting to collaborate with you.

Let’s get started.

The Agent Manager

Antigravity is built on top of the open-source Visual Studio Code (VS Code) platform, but it transforms the experience by shifting the focus from traditional text editing to intelligent agent coordination. Instead of a single workspace, Antigravity introduces two main views: the Editor and the Agent Manager. This design reflects the difference between doing the work yourself and overseeing how the work gets done.

Agent Manager View: Your Mission Control

When you open Antigravity, you’re not presented with a file explorer like in most IDEs. Instead, the first thing you typically see is the Agent Manager, your central hub for monitoring and directing the Agent’s activities, as illustrated below:

This interface functions as a centralized Mission Control hub, built for overseeing complex workflows. It gives developers the ability to launch, track, and collaborate with several agents running in parallel, each handling different tasks or parts of the project independently.

In this environment, the developer operates more like a systems designer, setting broad goals rather than performing detailed edits. Examples of such high-level instructions include:

• Revamp the authentication system
• Refresh or reorganize the dependency graph
• Create a full test suite for the billing API
  
  

Each instruction initializes its own agent, as shown in the diagram. The dashboard then presents a clear view of all active agents, showing their progress, the artifacts they generate, such as plans, outputs, and code changes, and any actions awaiting your confirmation.

This model solves a major drawback of earlier IDEs built around chatbot interactions, where work happened in a single-threaded, back-and-forth manner. In those setups, developers had to wait for one AI response to complete before issuing another request. Antigravity removes that constraint: you can assign multiple agents to multiple problems at once, significantly boosting development velocity.

When you click Next, you’ll be able to open a Workspace and continue setting up your environment.

Think of Workspace as you know from VS Code and you will be done. So we can open up a local folder by clicking on the button and then selecting a folder to start with. In our case,  I had a folder in my home folder named learninggravity, and I selected that. You can use a completely different folder.

Once you complete this step, you will be in the Agent Manager window, which is shown below:

Take a moment to review both the Planning and Model Selection dropdowns. The Model Selection menu lets you choose which available model your Agent should work with. The current list of models is shown below:

Similarly, we find that the Agent is going to be in a default Planning mode. But we can also go for the Fast mode.

Let’s look at what the documentation says on this:

Planning Mode

In this mode, the Agent takes time to think before acting. It’s ideal for tasks that require deeper reasoning, extensive research, or multi-step coordination. The Agent breaks the work into structured task groups, generates detailed artifacts, and performs thorough analysis to ensure higher-quality results. You can expect significantly more output and reasoning when using Planning mode.

Fast Mode

Fast mode instructs the Agent to act immediately without extended planning. It’s best suited for quick, straightforward tasks—like renaming variables, running simple shell commands, or making small, localized updates. This mode prioritizes speed and is appropriate when the task is simple enough that quality risks are minimal.

For now, we’ll stick with the default settings. Keep in mind that Gemini 3 Pro model usage is subject to limited free quotas at launch, so you may see notifications if your quota runs out.

Next: Understanding the Agent Manager

Let’s take a moment to explore the Agent Manager window. This will help you understand its core components, how navigation works in Antigravity, and how to interact effectively with the Agent system. The Agent Manager interface

1. Inbox: Think of this as a way to track all your conversations in one place. As you send Agents off on their tasks, these will appear in the Inbox, and you can click on the Inbox to get a list of all the current conversations. Tapping on any of the conversations will lead you to all the messages that have been exchanged, the status of the tasks, what the Agent has produced or even if it is waiting for an approval from your side on the tasks. This is a great way to come back later to a previous task that you were working on. A very handy feature.
2. Start Conversation: Click on this to begin a new conversation. This will directly lead you to the input where it says Ask anything.
3. Workspaces: We mentioned Workspaces and that you can work across any workspace that you want. You can add more workspaces at any time and can select any workspace while starting the conversation.
4. Playground: This is a great way by which you can simply start a conversation with the agent, and then, if you’d like to convert that into a workspace, where you have stricter control over the files, etc. Think of this as a scratch area.
5. Editor View: So far, we are in the Agent Manager view. You can switch at any time to the Editor view, if you’d like. This will show you your workspace folder and any files generated. You can directly edit the files there, or even provide inline guidance, commands in the editor, so that the Agent can do something or change as per your modified recommendations/instructions. We will cover the Editor View in detail in a later section.
6. Browser: Finally, we come to one of the clear differentiators that makes Antigravity very powerful, and that is its close integration with the Chrome browser. Let’s get going with setting up the Browser in the next section.
   
   

WANT TO GET A DETAILED TUTORIAL OF GOOGLE ANTIGRAVITY?

Watch this video!!!