Browser use has been rolling out in Antigravity, closing the loop between researching, writing code, and verifying it live. You choose what to log into, downloads pause until you approve them, and the agent can use the browser to run multiple tasks at once. Give it a try today.

Try it in Antigravity today

Will Gemini and Antigravity merge at some point? Please. 😉

SHIPPING TEAM

FINALLLLLYYYYY!! Antigravity might be the most powerful freaking harness ever made

Funny because antigravity had browser use a year and a half ago. What happened

Hydrogen when?

the approval checkpoints are interesting, especially once agents start interacting with real people. we ran a dating app experiment with 5 models recently and it raised some pretty uncomfortable questions about disclosure and trust

to check if it's reached you yet: settings > browser > browser actuation rules. it says it runs in the built-in browser when it's on, and requires google chrome when it isn't (rollout not there yet, a business or enterprise account, or an app older than 2.30)

The future looks promising. Keep building and keep raising the standards! 🌱

Without sufficient AI quota, all those extra features are just window dressing. Look at your "Gemini 4 Pro"—it can’t even complete a single delegated task before the quota runs out. Have you ever stopped to consider that this is what users actually care about?

we need custom context windows secelt but great update

What ever happened to computer use?

checking it now

Antigravity didn't have browser use ALL THIS TIME??????

@askmycindy 解释一下

😂😂 Wild that every other lab has had this since the beginning of the year

Good things appearing

Browser use in Antigravity closing the loop between research and code is exactly what I wanted, the agent verifying its own work in real time is the part that finally clicks

REMOTE FROM PHONE WHEN?

Thank you, Anthropic ！😏

finally I can watch the AI break my code in real time

does each parallel task get its own browser session? two agents sharing a logged-in tab sounds like a new kind of merge conflict

I remember Antigravity doing this same thing last year or maybe it was in AI studio.

Please implement a Pick & Polish feature similar to the one in GitHub Copilot. It's incredibly useful

@grok antigravity 不死之前就有浏览器功能吗？梳理总结一下现在加了什么功能，都有什么玩法？使用这个浏览器和使用 browser-use 以及 ego-lite 控制浏览器有什么区别？ 做一些深入对比

Verifying live in the browser is the part that matters most. If the agent can check its own work against the real page, the loop gets tighter and the review burden shifts from reading code to judging outcomes.

Always a year behind✌🏻😭

you need to just merge everything, main app should be chrome with antigravity as side panel with the option to open gemini integrated on the same sidepanel or another one, so you can have both opened, with obvious computer use as well. I have everything opened right now...

@grok busa tolong jelaskan?

Very impressed by the UI of the Swarm Mode especially with Knowledge Work applied. Keep cooking guys 💙

为什么我的ultrl会员，第三方额度周额有50%，5小时额度100%，你一次任务直接把两者全部消耗完毕了？而且，周额消耗的比5小时的还快？

downloads pause until i approve them, finally a browser with more restraint than me at 2am on a red candle