1 of 8

By: Patryk Dolaciński, Mateusz Gancarek & Piotr Sikorski

2 of 8

Summary web service application

Summar’z3 aims to save your time by transcribing audio files and providing them to openAI to transcribe and summarize them.

The user can both upload the audio file in mp3 format from his pc or provide a youtube url (provided the owner of the video allows downloading).

3 of 8

Main page layout

Our application was written in python. We have used flask for server side

To ensure that summary is precise enough we are dividing the transcript into many pieces that we then feed to GPT piece by piece to process it

4 of 8

Example results

Our application was written in python. We have used flask for server side

To ensure that summary is precise enough we are dividing the transcript into many pieces that we then feed to GPT piece by piece to process it

5 of 8

Access

The page can be temporarily accessed under:�http://88.199.63.212:2137/

The source code is available at:�https://github.com/AR621/Summ3r-y

6 of 8

What we used - AI

For transcribing the audio we have used OpenAI Whisper API

For summarizing we have used OpenAi GPT API with “text_curie_001” model

7 of 8

What we used - TECH

Our application was written in python 3.10. We have used flask for server side operations.

To ensure that summary is precise enough we are dividing the transcript into smaller pieces that we then feed to GPT piece by piece to process it.

8 of 8

The END

Thank you for your attention, and for the pleasure of taking part in this Hackathon. :)