In this episode, we use a small AI model to check the tone of a message in real time on the client's browser. There's no need to send messages back to the server or to an API because it's all done on the client's browser with the Web GPU. It is small enough to work on mobile devices and desktop computers.
Continuing from the previous episode, we look at adding the LLM insights functionality to examine the error logs. We start simple with the information captured in the error logs, but progress to provide code snippet context.