realtime-microphone-transcription

Realtime Microphone Transcription

Gradio web interface for real-time speech-to-text transcription from your microphone.

Demo

Features

• Web UI for live transcription powered by Gradio
• Stream microphone audio through WebSocket in real-time
• Display live transcription results as you speak

Requirements

Base dependencies are installed via the root requirements.txt. See the main README for setup. Add SMALLEST_API_KEY to your .env.

Extra dependencies:

uv pip install -r requirements.txt

This installs numpy (gradio is already in the root install).

Usage

uv run app.py

Open http://localhost:7860 in your browser.

1. Click the microphone button to start recording
2. Speak — transcription appears in real-time
3. Click stop to end the session

Recommended Usage

• Visual web interface for live microphone transcription — great for demos and prototyping
• Quick testing of Pulse STT with real-time audio
• For a full voice assistant with LLM and TTS, Jarvis is recommended

How It Works

The Gradio app captures microphone audio, streams it via WebSocket to Pulse STT, and displays interim and final transcripts in real-time. The WebSocket connection is managed using websockets.sync.client for thread compatibility with Gradio.

API Reference

• Streaming Quickstart
• Pulse STT WebSocket API

Next Steps

• Jarvis Voice Assistant — Full assistant with wake word detection, LLM, and TTS
• Streaming Transcription — Stream audio files via WebSocket