Inspiration
We were interested in creating a tool that makes comprehensive notes for service workers to increase accountability and challenge bias.
What it does
Our app takes an audio recording, transcribes it, and creates comprehensive summary lists about the information.
Frontend: We used Streamlit as a frontend tool to create our website and input generated summary lists into a viewable format.
Backend: We used Python because of the whisper python library’s ability to generate .json files for our Gemini Prompting.
Machine Learning: We used Gemini API to summarize transcriptions of police bodycam footage by sending raw text data along with a detailed system instruction for concise, actionable summaries. The API processes the content and outputs a summary, which is then saved into a text file for further use.
How we built it
Our product uses Python Whisper to discern from multiple people in a conversation and translates that text into a file. That file is prompted for a Gemini API Wrapper. The AI then summarizes the information in an easily digestible way that ensures high accuracy of objective information for indexible notes.
Challenges we ran into
Importing multiple libraries and ensuring that each section of code was able to be passed from one library to the next proved to be difficult Creating applicable json files for the Gemini API
Accomplishments that we're proud of
Working on our first ever front end development of a project
What we learned
Google genai libraries Python Whisper Transcription coding Streamlit
Project demo coming soon!
Log in or sign up for Devpost to join the conversation.