Automatic-To-Do-List-Generator

audio-recording-box
webpage

Automatic To-Do-List Generator Web App

This project aims to create a web application using Flask that automates the process of generating to-do lists from recorded audio. It leverages various technologies to achieve this, including Record-RTC for audio recording within the browser, openai-whisper for audio transcription using Python, and LLama3 NLP model for generating actionable to-do items. The app will integrate with third-party productivity tools like Google Calendar to set calendar events based on generated tasks.

Features:

Audio Recording: Record audio directly within the browser using Record-RTC library.
Transcription: Transcribe recorded audio into text using openai-whisper library.
Task Generation: Use LLama3 NLP model to convert transcriptions into actionable to-do items.
Integration: Connect with external productivity applications (e.g., Google Calendar, Notion) to set reminders based on generated tasks.

Project Status:

This project is currently a Proof of Concept and requires extensive refinement and development to enhance functionality, reliability, and user experience.

This Flask application serves as a web-based transcription service for audio files. Here's a summary of its functionality:

Key Components:

Flask Routes:
- /: Renders the home page (index.html) where users can upload audio files.
- /transcribe (POST): Handles the transcription process. It receives base64-encoded audio data, decodes it, saves it as a WAV file, performs transcription, and returns the result.
Dependencies and Libraries:
- Flask: Framework for web application development.
- base64, os: Standard Python libraries for file handling and base64 decoding.
- openai-whisper: Audio to Text converter to handle the transcription process.
- llama3: NLP Model to extract information from the transcribed text, to create accurate event information.

Deployment:

The application is configured to run locally (app.run(debug=True)). Deployment to a production environment would typically involve configuring a production server (e.g., Gunicorn) and ensuring appropriate security measures.

Future Considerations:

-Further Development: Extending the integration with more Productivity Applications (e.g Notion, Outlook)

Error Handling: Enhance error handling to provide informative responses to users.
Security: Implement secure file handling and consider authentication for user access.
Performance: Optimize for larger audio files and concurrent requests.

Summary:

This Flask app allows users to upload audio files for transcription. It employs base64 encoding to handle audio data sent from the client-side. Upon receiving the audio data, it decodes and saves it to a temporary WAV file. The transcribe endpoint processes this audio file using a custom module (generate_todo_from_audio from modules.py) for transcription. After transcription, the temporary file is deleted, and the transcription result is returned to the client as a JSON response.

The Google Calendar integration is currently a WIP, due to issues with the API Authentication. However, we do not perceive this to be a huge obstacle in the software development of our product.