Gemini in Chrome reaches Android users in the US with Auto Browse for subscribers
Google says Gemini in Chrome is now available to all Android users in the US. AI Pro and AI Ultra subscribers also get Auto Browse for selected multi-step web tasks, with confirmation required for some sensitive actions.
The short version
Google says Gemini in Chrome is now available to all Android users in the United States. The browser assistant can summarize long articles, answer questions about a page, connect with Google Calendar and Keep, and customize or generate images with Nano Banana. Google's announcement is the source for the rollout and feature list.
AI Pro and AI Ultra subscribers on Android in the United States also get Auto Browse. Google describes it as an agentic capability for tasks such as booking parking, updating a recurring online order, or organizing travel. Google's official post says Auto Browse asks for confirmation before some sensitive tasks.
What is new on Android
Gemini in Chrome opens from the Gemini icon or the three-dot menu in the top bar of an Android device. It can work with the page currently open, so the user does not need to switch to a separate assistant app for every question. Google documents the entry points and page features.
The rollout is for Android users in the US. This is a regional availability statement, not a global Android launch. Teams with users elsewhere should verify the current country and account requirements before communicating availability. Google's announcement specifies the US scope.
What Auto Browse can do
Auto Browse is designed for multi-step web errands. Google's examples include booking parking for an event, changing a recurring online order, and organizing travel. These examples describe intended use cases from Google, not a guarantee that every website or task will work. Source: Google.
The feature is limited to AI Pro and AI Ultra subscribers on Android in the US, according to Google. The announcement does not describe a separate usage price for Auto Browse, so the relevant commercial condition is the applicable subscription and the user's region. Source: Google.
Safety and confirmation
Google says the features are designed with security controls, including model training to detect known threats such as prompt injection. Google also says Auto Browse asks for confirmation before completing some sensitive tasks. The announcement links to Google's agent security explanation.
The confirmation boundary is important for teams. A browser agent can prepare a plan and navigate several pages, but a purchase, account change, or other sensitive action may still require a person to review the final step. The exact controls depend on the product and site, so teams should test with non-production accounts first.
How it compares with other action-taking assistants
Gemini in Chrome uses the browser as its working surface. Claude Cowork uses connected files and tools across desktop, web, and mobile, while Claude's Google Workspace connector can send, reply to, and forward Gmail messages with approval controls. Anthropic's connector documentation and Cowork documentation describe those capabilities.
The comparison is not only about which assistant can click more pages. Teams should compare the system of record, permission scope, review point, failure recovery, and audit trail. A browser task that succeeds but leaves no shared record may still create operational work for the next person.
What teams should test first
- Use a test account with no sensitive saved payment or identity data.
- Start with read-only research and reversible updates.
- Record which pages were visited and which actions were proposed.
- Measure how often the agent asks for clarification or human confirmation.
- Confirm what happens when a website changes its layout or blocks automation.
What this means for AI-native workspaces
A browser agent becomes easier to govern when the request, source links, permissions, decision, and result stay together. Kylon gives teams channels, files, agents, connections, skills, and recurring workflows in one workspace. A team can keep a browser task's context in a channel, assign review to a person, and preserve the completed result for later work.
Bottom line
Gemini in Chrome is now available to all Android users in the US, while Auto Browse extends multi-step web work to AI Pro and AI Ultra subscribers in that region. The release makes browser context more actionable, but the value for teams will depend on regional access, confirmation design, and the quality of the record left after the task. Google's official announcement remains the primary source.
Hire the AI agent team that runs your entire business.
