Here's a list of really cool Gemini-powered projects, demos and hacks that we liked. We hope that they will give you some ideas on what to build with Gemini next.
You should also have a look at our Google-maintained quickstarts and examples.
Those projects are ordered in an anti-chronological order and are the work of their authors, who are are not necessarily affiliated with Google.
The context caching feature for Gemini is really neat!
— elvis (@omarsar0) June 18, 2024
As you all know, I write a lot and document AI research progress in lots of places. But I also tend to forget research findings due to the vast amount of papers I read so I wanted to try whether context caching can help me… pic.twitter.com/7Q7G2RapFD
Gemini analyze the math drawn then computes an answers as well as where to display it. It can even recognize variables, functions, and solve brain teasers.
finally we can do maths pic.twitter.com/SoECzN1Aom
— tldraw (@tldraw) June 11, 2024
a b test pic.twitter.com/kBzxvGL91t
— tldraw (@tldraw) June 12, 2024
Bee maths pic.twitter.com/NPzfWl6UrS
— tldraw (@tldraw) June 15, 2024
maths is solved https://t.co/kgbc1soUhr pic.twitter.com/g9QwcICglB
— tldraw (@tldraw) June 13, 2024
[Vision] Gemini spatial example
Image analysis and labelling using gemini flash.
See also:
Alright Gemini's vision capabilities are really good. All I sent to Gemini was the photo and some instructions — the SVG on top was 100% generated by Gemini, including the coordinates etc. pic.twitter.com/wwv4a6maEa
— Blixt (@blixt) June 7, 2024
Gemini detects which application is used and which action is used in the app.
I am building a system to teach Gemini Flash new skills by observing the actions I take on my computer so it can repeat them.
— Pietro Schirano (@skirano) June 6, 2024
I found a way to achieve 100% accuracy in Gemini's ability to recognize what I am clicking.
The trick?
Draw a blue circle where the click happens. 🔵 pic.twitter.com/sxj6BcXSta
A Gemini API integration that labels your drawings.
this demo may seem un-flashy but its actually huge for letting models interact better with a canvas, and also potentially massive for accessibility use cases https://t.co/58MFuLQ3kP
— Lu wilson 🏳️🌈🏳️🌈 (@TodePond) June 4, 2024
[Vision and Code generation] Gemini UI to Code Streamlit App
Generates an HTML page from an image.
Introducing Gemini UI to Code 🧑💻
— Pietro Schirano (@skirano) June 3, 2024
A small but powerful app that uses an agentic framework to convert an image to code.
Perfect as a starter template or for learning new tricks.
Thanks to its improved Vision capabilities, Gemini 1.5 Pro is perfect for this task. 🪄 pic.twitter.com/QfNi0q4ygL
Gemini recognize the game, and the4 levels and provide advices when asked.
I built my own omni assistant using Gemini 1.5 Flash to guide me through Super Mario 64.
— Pietro Schirano (@skirano) May 21, 2024
Gemini can see what I do on my screen and communicate with me in real time via voice, and thanks to the long 1M context, it has a memory of everything we do together.
Incredible. pic.twitter.com/doTngufjFL