Google has added a built-in browser and PDF viewer to its Antigravity app, allowing developers to open websites, inspect local web applications and read PDF documents without switching to external applications. Both features appear in the app’s side panel, alongside the ongoing AI-agent conversation.
The additions arrived with Antigravity App version 2.28.0, released on October 27, 2026. The update also introduced support for attaching files of any type up to 20MB, more detailed code-diff highlighting and a shortcut for reopening closed side-panel tabs. Google’s changelog says updates are distributed gradually, so availability may vary between users.
Key Takeaways
- Google Antigravity App version 2.28.0 introduced a built-in browser and PDF viewer.
- Websites now open in a browser tab inside the side panel.
- Links opened from conversations use the built-in browser by default.
- Users can read PDF documents without leaving the Antigravity application.
- The built-in browser supports navigation, separate sign-ins and website security indicators.
- Users can select text from websites and PDFs to send comments or quoted passages to the AI agent.
- Antigravity now supports attachments of any file type up to 20MB.
- The update also improves code-diff highlighting and allows users to reopen closed side-panel tabs.
What Is the New Built-In Browser?
The built-in browser is designed to bring web browsing directly into the Antigravity workflow.
Previously, developers often had to switch between their AI coding environment and an external browser to inspect websites, read documentation or test applications.
With the new feature, web pages can open in a dedicated browser tab inside Antigravity’s side panel.
This allows users to keep their AI conversation visible while viewing a website alongside it.
For example, a developer building a website can open the running application in the side panel, inspect its appearance and then ask the AI agent to make changes.
The browser also supports local development addresses, including applications running on a computer through URLs such as http://localhost:3000.
This is particularly useful for developers who want to test a locally running website without constantly switching windows.
Google also designed the side panel to widen automatically when a web page opens, helping websites display their desktop layouts rather than switching to narrow mobile versions.
How the Browser Works Inside Antigravity
The browser includes several familiar navigation features.
| Feature | Function |
|---|---|
| Address bar | Enter a website address or search query |
| Back and Forward | Navigate between previously opened pages |
| Reload | Refresh the current website |
| Browser tabs | Keep multiple websites open in the side panel |
| External browser | Open a website in the default browser |
| Security indicator | Check connection security and certificate details |
One important detail concerns website sign-ins.
The built-in browser maintains its own cookies and login sessions, separate from the user’s Chrome profile.
This means websites opened inside Antigravity do not automatically inherit the user’s existing Chrome sign-in sessions.
The separation can help keep browsing contexts distinct, although users may need to sign in to websites again inside the built-in browser.
Google also displays whether a connection is secure, including certificate details, or shows a warning when a website is not secure.
AI Agents Can Use Website Content as Context
One of the most useful additions is the ability to share selected website content with the AI agent.
Users can select text on a web page and choose the Comment or Quote option.
The selected passage, along with the page URL, can then be included in the next message to the agent.
This creates a more direct workflow for developers and researchers.
For example, a developer reading API documentation can select a particular requirement and ask the agent to implement it in a project.
Similarly, someone reviewing a competitor’s website can quote a specific section and ask the agent to analyse its design or functionality.
Google also documents a /browser command that uses a browser subagent to browse and test pages through Google Chrome.
This is distinct from simply displaying a page in the built-in side-panel browser. It allows the agent to carry out browser-based tasks and testing.
Antigravity Adds a Built-In PDF Viewer
The second major addition is a native PDF viewer.
PDF files can now open directly inside Antigravity, whether they are opened through the file viewer or generated as an artifact.
This removes the need to open a separate PDF application when reviewing technical documents, research papers or project specifications.
The viewer includes several document-navigation features:
- Page thumbnails for navigating longer documents.
- Page navigation controls.
- Zoom controls.
- Fit-to-page and fit-to-width viewing options.
- Text selection and quoting.
- Options to search, download or copy the file path.
Users can also select text from a PDF and send it to the AI agent as a comment or quoted passage.
For example, a developer working with a technical specification can highlight an API requirement in a PDF and ask Antigravity to use that requirement while building an application.
The combination of a PDF viewer and website browser makes it easier to use external documentation as context without leaving the development environment.
Antigravity Now Supports More File Attachments
The version 2.28.0 update also expands the types of files users can attach to conversations.
Previously, some file formats were rejected because they were not supported as conventional attachments.
The updated application allows files of any type, including videos, design files and programs, up to 20MB.
Files that were previously rejected can now be attached as files, with the agent reading them from where they are saved.
This could help developers share a wider range of project materials with AI agents.
For example, a user could attach a design file, a video showing a software bug or a program file and ask the agent to examine it.
However, this does not mean that every file format can be fully understood or previewed inside the application. The agent’s ability to interpret an attachment depends on the file’s contents and supported processing capabilities.
Code Review and Side-Panel Improvements
Google has included several additional improvements in the same release.
One notable change is more detailed code-diff highlighting.
When an AI agent edits a file, Antigravity can now highlight the exact words and characters that changed within a line. This also applies to edits spanning multiple lines.
The improvement makes it easier to identify small modifications when reviewing AI-generated code.
The update also introduces a keyboard shortcut for reopening recently closed side-panel tabs.
| Operating system | Shortcut |
|---|---|
| Windows | Ctrl + Shift + T |
| Linux | Ctrl + Shift + T |
| macOS | Cmd + Shift + T |
The shortcut restores closed tabs at their previous position and scroll point.
Google has also added new status icons to indicate whether an agent is working, running a background command or waiting for the user’s response.
These improvements make it easier to understand what an AI agent is doing during longer development tasks.
How to Use the New Browser and PDF Viewer
Users with the updated Antigravity app can access the features through the side panel.
To open the built-in browser:
- Open Antigravity.
- Click the plus (+) button in the side panel.
- Select Browser.
- Enter a website address or search query.
- Use the browser alongside the AI conversation.
Links opened from agent responses also open in the built-in browser by default.
Users who prefer an external browser can change the Open Links In setting under Settings, then General.
To open a PDF:
- Click the plus (+) button in the side panel.
- Select Open File.
- Search for the PDF in the project.
- Open it in the built-in viewer.
- Select text and use Comment or Quote to share relevant passages with the agent.
PDF links opened on the web can also be viewed through the built-in browser.
Which Antigravity Version Includes These Features?
Google introduced the browser and PDF viewer in Antigravity App version 2.28.0 on October 27, 2026.
Subsequent updates added further improvements. Version 2.28.1 introduced a conversation navigator and additional information about third-party plugins, while version 2.28.2 expanded Marketplace functionality.
Google says new versions roll out gradually and may take several days to reach all users.
Developers can check the official Antigravity changelog to see the latest release and its improvements.
The Bigger Picture
The addition of a browser and PDF viewer reflects Google’s effort to make Antigravity a more complete environment for AI-assisted development.
Developers frequently move between coding tools, browsers, documentation, design files and PDF specifications. Bringing these activities into one interface can reduce context switching and make it easier to provide AI agents with relevant information.
The ability to select passages from web pages and PDFs is particularly important because it connects external information directly to the agent’s workflow.
Rather than describing a requirement manually, users can point the agent to the exact information they want it to consider.
Looking Ahead
Google is continuing to expand Antigravity through improvements to its browser, side panel, AI agents and developer tools. Future updates could make the connection between browsing, document research and software development even more seamless, although the company has not announced a specific roadmap for these features.
For developers, the immediate benefit is a more integrated workspace. They can inspect websites, read technical PDFs, attach additional file formats and review AI-generated code changes without constantly moving between separate applications.
Get the day’s top stories in your inbox
One concise email. No spam, unsubscribe anytime.
