Speech To Text🔈 API written in python using flask micro-architecture as backend.

💻Technology Stack

1. Frontend : HTML, BOOTSTRAP, Recorder.js, AudioDisplay.js
2. Backend : Flask
3. Speech Recognition : (Includes Two MODELS : DeepSpeech, CMU Sphinx)
4. Deployment: wsgi, aws
5. Endpoints : "/generate_transcript" "/download_transcript" /generate_transcript: Accepts POST request parameter "file" with audio form data from recorder.js /download_transcript: Accepts GET and downlods transcript saved in output.txt

⚒Testing

Included 3 Audio files in Files/Audio/test_audio/

Transcript is stored in Files/Transcript/output.txt

⚡️Run

chmod +x run_me.sh && ./run_me.sh

(This script installs all the dependencies for this project and runs the flask application)

Note : This project is tested on python 3.6 on a mac running MAC os Catalina.

Manual Run can be done through running aigalore_mainfile.py. No arguments needed.

📁Sample Output

your power is sufficient i said
Total Recognised Words:6
Words Per Minute:169.81132075471697
Total Filler Words:3

📭About me

---

website: govardhanchitrda.com