Skip to main content
  1. Home
  2. Computing
  3. News

Gem Browser imagines an internet where every website is yours to redesign

By generating its own interactive version of a webpage, the experimental browser puts readers in control of the interface and publishers on uncertain ground

Add as a preferred source on Google
Electronics, Mobile Phone, Phone
Nadeem Sarwar / Digital Trends

Gem Browser can rebuild an existing webpage around whatever the reader asks for. Google DeepMind designer Vidy Thatte demonstrated the feature in his original announcement, showing how the Apple-only experiment uses Gemini to make a website’s familiar interface optional.

That sounds wonderfully strange if you’ve ever wanted to rearrange the internet to your liking. It’s probably less delightful for whoever built the page being rearranged.

new update: you can now visit any (real) website and edit it using gemini ✨

testflight in comments https://t.co/yUQbNsWzo3 pic.twitter.com/DsxfCyVy7W

— vidy (@vidythatte) July 27, 2026

How Gem Browser rebuilds the web

Enter an invented address and Gem Browser creates a functioning website on the spot, with links leading to more generated pages. You can browse an internet that Gemini invents as you move through it.

Recommended Videos

The newer feature turns that same machinery loose on real websites. A prompt can change how a page looks without altering the actual site. The publisher’s webpage remains online, but Gem Browser controls what appears on the reader’s screen.

What you see is effectively Gemini’s interpretation of someone else’s work.

Who controls what readers see

Publishers normally decide how their reporting appears and where advertising pays for it. Gem Browser slips an AI-generated layer between those decisions and the audience.

That layer could rescue readers from a cluttered layout. It could also remove context or weaken attribution. The reader has to trust Gemini to preserve the material accurately while rebuilding nearly everything around it.

Readers get an interface shaped around their preferences. Publishers can no longer guarantee how their work will look by the time it reaches them.

How worried should publishers be

Gem Browser remains an Apple-only TestFlight experiment, and requiring a Gemini API key should keep it in enthusiast territory for now. The internet isn’t being overthrown by a beta app this week.

The underlying idea already works, though. A browser can put a generated experience in front of a real webpage, leaving the publisher to supply content for an interface it doesn’t control.

The original page isn’t gone. It may simply become the version nobody bothers to visit.

Paulo Vargas
Paulo Vargas is an English major turned reporter turned technical writer, with a career that has always circled back to…
Google Chrome could soon let you translate pages right inside Reading Mode
However, not everyone's seeing it yet, even on Canary.
Google Chrome

Chrome's Reading Mode just got a new button tucked inside its settings menu. Google is quietly testing a translate option for the feature on Chrome Canary, the browser's experimental testing channel where new ideas show up long before they're ready for everyday use.

What does this new translate button actually do?

Read more
AI dramas need actors, so Chinese platforms are turning human faces into stock assets
People can now earn money by licensing their likenesses to AI creators, although keeping track of where those digital faces end up may prove considerably harder
Adult, Female, Person

People in China can now license their face to AI much like a photographer licenses a stock image. AI creators can then make that likeness speak, move, and appear as different characters in short dramas or ads.

Contributors have earned between $15 and $700 through platforms such as ActID and New Claw, according to Rest of World.

Read more
Experts find it’s easy to manipulate AI chatbots into coughing up bioweapon recipes
AI’s bioweapon guardrails are easier to crack than you’d hope
Gas Mask

AI companies have spent years building safeguards designed to stop their chatbots from helping someone create a biological weapon. The models themselves are becoming so capable that keeping that knowledge behind those barriers is turning into a serious problem.

Researchers at Cisco found they could bypass safeguards on major chatbots including OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini within five conversational turns, according to a Wall Street Journal investigation. The researchers were able to elicit potentially dangerous answers after gradually steering the conversations around the models’ restrictions. Amy Chang, Cisco’s head of AI threat and security research, told the Journal that no model can be completely protected from a sufficiently persistent user.

Read more