Loading video...
Video Failed to Load
Antigravity can do more than ship code and you don’t even have to leave your editor. In this demo, the agent reads a blog post, extracts the core narrative, and builds a Google Slides deck from scratch, handling the research and initial build for you.
361,739 views • 8 months ago •via X (Twitter)
34 Comments

Google may have designed Antigravity with coding in mind, but I’ve pushed its boundaries to create a truly limitless digital assistant. From visual design and complex coding to Kurdish history research, legal consultancy, and even patent applications, it’s become my ultimate partner. I’ve used it for everything from app development and document scanning to restoring two damaged books. Now, my entire digital workflow, including social media and book projects, runs solely on Antigravity. The possibilities are truly endless!

Yes just browsing feature is too slow, it need to be speed up

Increase the limits so we can make the best out of agy!

Makes more sense to add this to google slides itself as a button - “start draft from link” for example than have such a work around

I am spending my whole day with Antigravity ♥️

Too bad the quotas are getting ridiculously low, day after day

Do you know how much memory it eats on PC ?I got 128ram i9 intel heating hup like a heater last night!! This needs a fix

I cant wait for more tutorials on interesting ways to use it and its capabilities efficiently

It is amazing to use @antigravity as a pro user. Everyday I am experimenting new stuff from my LaTeX projects to building open source tools and keeping my blog alive. I am so happy I have subscribed to @GoogleAI pro and I invite whoever is thinking about it to try that out!

Ngl, watching this demo got me hyped — agent extracts narrative, researches, spits out ready Slides. For X growth peeps: Imagine prompting it with "turn this viral thread into a pitch deck for sponsors" or "repurpose my analytics post into investor slides". Time saved = more posts = more impressions. Algo loves consistent value. What's the biggest pain point this could fix for your workflow?

Antigravity needs to be studied

IShowSpeed

Ok yes, but what agentic use cases like this where there's not already a more capable tool in Gemini to do it? I'm so excited for agentic use from Antigravity but there's no documentation of what all it can accomplish.

Good move! Antigravity could become a one-stop shop for developers.

I'm on Gemini Pro, i think the limits are strict now , 1 hour slow usage and it's gone.

Cool

Please fix the login issue :( I can't use antigravity with my Google One ai plus account from days :( (Yes I already open several tickets but no response)

Turning long-form thinking into a ready-to-present slide deck inside the editor is powerful; curious how it handles nuance versus just summarizing content.

the first pass looks solid. curious about the ceiling though — can it get to "ready to present" quality or more like a good starting point you'd still polish manually?

Unlike ChatGPT or Claude, which might give you an outline for a deck, Antigravity acts as an orchestrator. It creates an Implementation Plan and a Task List (Artifacts) that you can audit and comment on—like a manager reviewing a junior's work—before it executes the research and file creation.

i'm very bullish on antigravity

It's more of a workspace for everything. And please fix the subagent that controls browser it's too slow.

Automating slide decks is a huge time saver feature

Can you increase opus limits please :)

help please fix this i am in the uk, THE UK UNITED KINGDOM ITS SUPPORTED.

Make it available in Andorra

good

Antigravity is Claude Code with a UI

it can not read echarts document website ☕️

But I still feel that kiro ide is miles ahead of Antigravity, even though I use both , I feel kiro is much much better and also I avoid using Gemini models as I only use Claude models for all my coding work.

Is this video sped up by like 10x? Why can’t the browser function actually be this fast?

Why from the last 2 days, after every prompt I see an agent execution error....

this is why AI agents will win. not the chatbot UX, not the voice interface, but the invisible background execution layer that just handles the tedious primitives while you stay in flow state. abstraction gradients > UI paradigms

You guys are basically building a closed source Co(lab) that only works with gemini models
