PulseCue — screenshot 1
PulseCue — screenshot 2

PulseCue

Find the Moment. Skip the Search.

Codex Community Hackathon - Pune

Additional info

How was your experience building with Codex?

CODEX is really good at coming up with detailed approaches that work, Found it to be a little slow, but that is okay considering you do not have to retry most of the time, also could be a prompting issue

Describe your experience using Loops House as the hackathon platform. What worked well, what challenges (if any) did you face, and what improvements would you like to see?

Really great experience

Tell us about your overall experience at Codex Community Hackathon Pune.

Really great

What could Codex Community improve to create a better experience for participants?

Probably provide more credits as a lot of folks ran out halfway through the hackathon :)

Team

2 members
  • AM

    Amey Kulkarni

    Team Member
  • AY

    Ayush Chanekar

    Owner

Overview

PulseCue is an RAG based search for youtube videos that uses LLM agents (chatgpt-o4) and OpenAI Whisper framework to search a particular and most relevant section of any given set of videos that the user wants to select to be trained upon. Users can select topics for the videos they want to search and PulseCue will use yt-dlp and youtube's official transcript api to understand the video content. Using embeddings and the power of openai's o4 model we can perform parallel semantic searches using LLM agents.