Prompt API

probably the simplest way to use AI in our browser. it uses gemini nano model. available on chrome version 140 and above (according to docs).
you can read more about prompt api here

If it is available. Great!

Now you can use it like this:

Chat with Gemini Nano Guide on how to use it

What if it is not available?

Then here are few things you can try that worked for me, it is not exactly documented in the docs. may or may not work for you.
My machine: Mac OS Tahoe: 26.0, Chrome: 142.0.7444.60 (Official Build) (arm64)

you have to be on chrome [windows, mac, linux] ofcourse. checked other chromium based browsers like brave, edge, etc.. they are not supported.
check your chrome version at chrome://settings/help should be 140 or above.
check if `Prompt API for Gemini Nano` flag is enabled at chrome://flags/#prompt-api-for-gemini-nano you can set it to enabled or enabled-multilangual
check if `Optimization Guide On-Device Model` flag is enabled at chrome://flags/#optimization-guide-on-device-model I have it set to `Enabled BypassPrefRequirement`
in chrome://components/ search for `Optimization Guide On Device Model` to download if you don't see this check previous steps and restart chrome. if you see it click on "Check for Update"

once you have done all the steps, restart chrome and check if it is available.