EchoTap

Real-time desktop transcription app that converts speech from any audio source into text locally on your machine.

What it does

EchoTap transcribes audio from any desktop source - system audio, microphone, or any application - into live text. Everything runs locally using the faster-whisper large-v3-turbo model with no external APIs or internet dependencies required.

Setup

Prerequisites

• Node.js 18+ and npm
• Python 3.8+ with pip
• Git

System Dependencies

macOS:

brew install portaudio

Ubuntu/Debian:

sudo apt-get update
sudo apt-get install portaudio19-dev python3-dev

Windows: No additional dependencies needed

Installation

git clone https://github.com/pmbstyle/EchoTap
cd EchoTap

npm run setup

npm run dev

GPU Acceleration (Optional, Recommended)

For 4-10x faster transcription, install PyTorch with CUDA support:

# Navigate to backend directory
cd backend

# Install CUDA-enabled PyTorch (requires NVIDIA GPU)
pip install torch --index-url https://download.pytorch.org/whl/cu121

# Verify GPU is detected
python -c "import torch; print('CUDA available:', torch.cuda.is_available())"

Note: If you don't have an NVIDIA GPU, the app will automatically use CPU (slower but works fine).

First Run

On first launch, EchoTap automatically downloads the faster-whisper large-v3-turbo model (~1.5GB). This enables completely offline transcription afterward.

Usage

1. Click the microphone button in the floating bar to start transcribing
2. Audio from your system or microphone will be converted to text in real-time
3. Click the transcript icon to view full transcription history
4. Click the archive icon to search through past sessions