Google is turning Chrome into an AI-first browser with three headline upgrades presented as one experience: Gemini in Chrome as a built-in assistant, a new AI mode in the omnibox for layered queries and follow-ups, and strengthened safety features that use machine learning to spot scams, reduce spammy prompts, and improve password and permission decisions. The promise is less tab shuffling, fewer repetitive searches, and quicker progress from question to action inside the browser you already use.
Rather than bolt on features, Chrome is designed to read context across pages and tabs so the assistance shows up where it saves time. Gemini can summarize long articles so you can decide whether to dive deeper, compare information across multiple open pages, and recall previously visited sites when you want to resume work without digging through history. The goal is to keep you in flow while the browser quietly handles the glue work between steps.
Here is everything new announced by Google:
-
Gemini in Chrome functions as a page and tab aware assistant that answers questions inline, summarizes content when helpful, and helps keep momentum across research, forms, and media without constant switching.
-
Agent-like workflows are coming that compress multi-step chores into guided sequences you approve at the end, such as filling forms or completing a routine checkout, so busywork shrinks without giving up control.
-
AI mode in the omnibox turns the address bar into a place for multi-part questions and follow-ups, providing an initial answer with relevant links so you can refine the prompt without losing your place.
-
Contextual suggestions appear in the omnibox based on the page you are viewing, like prompts about warranty or returns while you are on a product page, which shortens the distance from noticing a detail to acting on it.
-
Security upgrades use on-device and cloud AI to flag risky pages, cut down noisy notification requests, de-emphasize prompts that users typically deny, and strengthen password help including one-click changes on supported sites.
In practice, this shifts Chrome from a passive renderer to a helper that understands what you are trying to do and offers timely next steps instead of forcing you back to a blank search box. Summaries act as a triage layer, contextual prompts nudge the next question without breaking focus, and the assistant takes on predictable steps while preserving a human in the loop for final approval.
Rollout begins on desktop in the U.S., with English as the initial language. Gemini in Chrome arrives first on Mac and Windows, with Android and iOS support to follow. Contextual suggestions in the omnibox are available now in the U.S. in English, and AI mode begins later in September 2025 in English, with more countries and languages to follow in the weeks ahead. Some capabilities may require sign-in, apply only to select devices and regions, and be limited to users 18+, so the exact experience may vary.
Check out the introduction videos below:
