Streamlit Realtime Audio Recorder

A Streamlit component that records audio in real-time and automatically stops recording after a configurable silence period.

Installation

pip install streamlit-realtime-audio-recorder

Usage

from streamlit_realtime_audio_recorder import audio_recorder
import streamlit as st
import base64
import io

result = audio_recorder(
    interval=50,
    threshold=-60,
    silenceTimeout=200
)

if result:
    if result.get('status') == 'stopped':
        audio_data = result.get('audioData')
        if audio_data:
            audio_bytes = base64.b64decode(audio_data)
            audio_file = io.BytesIO(audio_bytes)
            st.audio(audio_file, format="audio/webm")
        else:
            pass
    elif result.get('error'):
            st.error(f"Error: {result.get('error')}")

Parameters

• interval (optional, default: 50): How often to check audio level in milliseconds
• threshold (optional, default: -60): Audio level threshold for speech detection in dB
• silenceTimeout (optional, default: 1500): Time in milliseconds to wait after silence before stopping recording
• play (optional, default: False): Whether to play the audio during recording

Demo

Check out the demo video showing the component in action:

How It Works

This component uses the Web Audio API and the hark library to detect speech and silence. When you click the microphone button:

1. The microphone starts recording
2. When you stop speaking (silence is detected), the recording automatically stops after a configurable timeout
3. The audio data is returned as base64-encoded data that you can play, save, or process

Use Cases

• Voice commands in Streamlit apps
• Speech-to-text integration
• Audio data collection
• Voice recording with automatic silence detection

License

MIT