neuralwrite

Inspiration

I am going to create the best text editor, neuralwrite. We live in a world where we mindlessly copy over the same "perfectly" generated texts from ChatGPT without ever reading it. neuralwrite is an AI-Powered Text editor where you will forever still be the main character. It will be the go-to editor for students, teachers, and analysts. (for now)

What it does (currently)

For context of the hackathon, neuralwrite uses Google ADK through the instant dashboard agent. The instant-dashboard agent is connected to BigQuery. The goal is to be an accurate insight and visualization resource for large datasets. Specifically, the agent takes the user's natural language question about the database, convert it to a structured JSON query plan, then a detailed SQL query, to a Chart.Js Visualization and AI - generated insights. All of these are kept as concise as possible, after getting such insights, the user can quickly paste it into their current writing.

How we built it

Obviously, the instant-dashboard backend structure is built using Google Agent Development Kit. However it is currently integrated into a larger project called neuralWrite where most of the app development magic happens. Instead of using the default web run, I felt it would be better suited as an API call for a larger application.

Challenges we ran into

I think one of the most challenging aspect is to find a way users will actually use it. I spent a lot of time trying to create the longest most complex multi agent system possible. Before realizing that no one has the attention span for a 35 second run time. Therefore, I pivoted towards building smaller feature that would be much suitable for larger applications.