Recipe with Voice Command

The idea of this app is to demonstrate voice to text and using that text generating recipes instantly for the user. Working on this concept, only the backend has been implemented using python Flask. Machine lerning model Whisper is used in the project to work with ffmpeg. For recipe suggestions and details, third party The MealDB is uded.

Features

• Voice to text convertion
• Generating recipe list suggestion based on the voice.
• Generating recipe details

Tech Stack

• Python 3.12.8
• flask 2.3.2
• pytorch 2.0.1
• torchvision 0.15.2
• torchaudio 2.0.2
• openai-whisper 20230314
• pydub 0.25.1
• ffmpeg-python 0.2.0

Installation

1. Clone the repository:

   git clone https://github.com/Tanjemul/recipe-with-voice-command.git
   

2. Navigate to the project directory: cd \recipe-with-voice-main

3. Install dependencies:

   pip install flask
   pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
   pip install openai-whisper pydub
   pip install pydub
   

On your local machine, install ffmpeg

4. Run the project

   set FLASK_APP=app
   flask run

On local machine, open browser and go to: http://127.0.0.1:5000

4. API Documentation: Only one end point is enough for this project:

      curl --location 'http://localhost:5000/transcribe' \
   --form 'file=@"/C:/Users/User_PC/Downloads/audio_file.mp3"'