Inline AI

Unleash the power of AI on any website! Enjoy instant prompts and in-browser AI at your fingertips, right in your side panel for smarter, faster and productive work session.

Comment

Inspiration

I use X and LinkedIn a lot and read online articles a lot find out that while replying and reading, there were some things I needed to open Gemini in another tab and keep on going back and forth for summary proofread, and creating a reply.

I found some extension solutions by giving you a chat window on any website, but they use Gemini, Claude, and GPT models and ask us to enter an API key which will cost money and your sensitive information might also go to these models.

So, I liked the idea of local, in-browser AI that processes things on the user's machine, and no API key is required. So non-technical users can also use these tools without any setup

What it does

This extension provides an AI chat window on every website inside the Chrome side panel. But one more powerful setting allows users to set a system prompt on the go. What should be the length of each response that AI gives.

Suppose you want to proofread. Instead of asking, again and again, that proofread this text.....

Users can set a powerful system prompt that will always do the best job of proofreading. Changing of system prompt happens on the go by just selecting the prompt from a dropdown.

How we built it

I used React and Tailwind to build this extension

Challenges we ran into

There were some times that I saw the error that the model had crashed I was not able to use the AI API

Accomplishments that we're proud of

Making this extension work in the side panel of Chrome and passing the text from the user's right click to the side panel was one task I was stuck but once it worked it created a very nice user experience

What we learned

This is the first Chrome extension I have made and that uses an in-built AI API. So I learned a lot about how the Chrome extension works and also about the extension API that Chrome provides. The docs were always helpful.

What's next for Inline AI

These are my top 3 ideas

I will add more models in the extension setting that users can choose from. For example, the user can choose between Gemini and Gemini-nano.

A setting in the Side panel when the writer or rewriter is chosen, the user can set and edit the sharedContext.

A button to take a screenshot of the browser tab and chat with it if any model is chosen that understands the image.

Built With

Share this project:

Updates