Recapify

Inspiration

In today's fast-paced world, students and professionals often struggle to keep up with lectures and meetings. Even if lectures or meetings are posted online, most professionals do not have the time to review the entire discussion. We wanted to create a tool that makes it easier to review important discussions without replaying hours of audio. Recapify was born from the idea of making note-taking more efficient and accessible.

What it does

Recapify is a web application that allows users to upload an MP3 file, transcribe the audio into text, and generate a concise summary of key points. Both the transcript and summary are displayed on the website. The application allows users to download a PDF with the summary that was created by clicking a link. The user has the option to send the PDF as an email attachment to a specified email address directly from the website. This makes it a great tool for students, professionals, and anyone who needs quick and accurate meeting or lecture notes.

How we built it

We developed Recapify using:

• Frontend: HTML, CSS for a clean, user-friendly interface
  
• Backend: Python with Flask to handle file requests, uploads and processing
• OpenAI's Whisper API: Used to accurately transcribe audio files into text
  
• Summarization Algorithm: Extracts key points from the transcription, using OpenAI prompts -reportlab library: Used to generate the PDF -Gmail API: Used to send email(s) with the PDF attached to a specified email address

Challenges we ran into

• Setting up OpenAI to allow for transcription of audio
• Handling file requests between the python script and html
• Generating PDFs with proper formatting consistently
• Implementing the functionality of the email attachment
• Integrating all functionalities together while minimizing errors

Accomplishments that we're proud of

• The functionality of the application and the successful implementation of all components
• Implementing an accurate transcription system that converts audio into text
• Allowing users to download summaries as PDFs and email them directly from the application
• Designing a clean webpage and making it work with the back-end such as the file requests.

What we learned

• How to work with speech recognition and text summarization APIs
• Implementing PDF generation and email-sending functionality
• How to efficiently develop a user-friendly webpage
• How to integrate front-end and back-end so that they work together

What's next for Recapify

• Enhancing the accuracy of transcriptions by integrating a more advanced AI model
• Adding a real-time transcription feature for live lectures and meetings
• Implementing multi-language support
• Improving the UI/UX for a smoother user experience
• Formulating PDFs with better formatting and organization of the notes

Built With

• css
• figma
• flask
• gmailapi
• html
• javascript
• openai
• python
• vscode
• whisper