Demos
Getting started: See what the toolkit does · Demos · Build your own project · Get an API key · Submit your project
Projects built with the toolkit, to show you what it can do and give you ideas. Each one links to its code.
A reading app, built step by step
This one-page reading app changes to fit what the reader asks for, such as “bigger text”, “I’m dyslexic” or “read the next paragraph”, and remembers it for next time. You can watch a developer build it, or read its code and tutorial on GitHub.
Read a text version of the video, or open the video on Google Drive.
A driving game that adapts to the player
A driving game changes to fit the person playing it, and remembers what they asked for. Tell it what you need in your own words, such as “I can’t see the screen” or “I get motion sick”. Someone who can’t see the screen can play by ear: a tone shows where the middle of the road is, and a voice warns of the curves ahead. You can also drive by voice, with words such as “go” or “left”.
It’s the second tutorial after the reading app, and it shows developers how to make the toolkit work in a new kind of app, a game instead of a web page. Running it needs a Gemini API key.
Browser extensions
These three Chrome extensions adapt the websites you visit. All three need Google Chrome on a computer. The install guide walks you through the first two, and the workshop tutorial has its own short steps.
- The original extension changes pages to suit a profile you pick, such as low vision. It can make text bigger, turn the page dark, stop moving content, describe pictures, add captions and rewrite complicated text in plainer words.
- The personalized extension first asks what you need, then remembers what helps and suggests changes as it learns. It can also build new adapters for you, and it works by voice.
- The workshop tutorial, as an extension puts the tutorial’s panel on real websites, such as Wikipedia, so you can try adapters and build new ones on the sites people actually use. Your profile and adapters go with you from site to site, and there’s no key or account to set up. If you’d rather not install anything, the same panel works on Redbird, a practice online store.
A browser harness
A browser harness lets an AI agent use your own Chrome browser for you. It applies your ability profile, such as bigger text or dark mode, to every page it opens, and switches on Chrome’s own live captions and image descriptions when your profile asks for them. It reads each page the way a screen reader does. In a chat, you can ask it to go to a page, search or do a task.
Guidance on smart glasses
An audio-first guide for smart glasses talks a person through a household task one step at a time and reads out what they’re looking at. You can switch between a standard mode and a toolkit mode on the same task, and hear how the toolkit changes the wording and the speech to suit the person’s profile.
Read a text version of the video, or open the video on Google Drive.