georgesung / whisper_asr_app

Streamlit app to record audio and run speech recognition (ASR)

Demo Whisper ASR app

Basic Streamlit app to record audio from user's microphone, run Automatic Speech Recognition (ASR) using distil-whisper-large, and show the transcribed text.

Pre-reqs

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Set up HTTPS (if needed)

If running beyond localhost, you may need HTTPS so the browser can access your microphone. Make note of the port the app is running on, e.g. 8501. Then set up an HTTPS proxy using https://github.com/suyashkumar/ssl-proxy/ (change the command below to match your OS):

./ssl-proxy-linux-amd64 -from 0.0.0.0:<external_https_port> -to 127.0.0.1:<internal_http_port>

E.g.

./ssl-proxy-linux-amd64 -from 0.0.0.0:8509 -to 127.0.0.1:8501

Where internal_http_port is the port the app is running on (e.g. 8501), external https port is the port to serve on (e.g. 8509 -> https://your_url:8509)

How to run NiceGUI app

This app uses NiceGUI

python ngui.py

How to run Streamlit app

This app uses https://github.com/stefanrmmr/streamlit-audio-recorder

streamlit run app.py

How to run Streamlit WebRTC app

This app uses https://github.com/whitphx/streamlit-webrtc

streamlit run app_webrtc.py