🗣️📝 VoiceWriting

VoiceWriting is a Whisper based Desktop application to write using your voice. It works in real-time and even fixes common grammatical, punctuation and other issues in your speech.

Give it a try by following the instructions below, or check out the demo 📹 below 👇🏻.

voice-writing-demo.mp4

🛠️ Easy Setup

1. Clone and install dependencies

git clone https://github.com/aviaryan/voice-writing-electron.git
yarn install
cp .env.copy .env

2. Populate .env file

1. Add the absolute path for whisper.cpp's stream executable in STREAM_EXECUTABLE_PATH. Learn more about stream.
2. Add the absolute path for GGML Whisper model in MODEL_PATH. Learn more about model.
3. Fill in the GROQ API Key. Learn more about Groq API key.

Example -

GROQ_API_KEY=gsk_abcdksk2kkskssd
STREAM_EXECUTABLE_PATH=/Users/john/dev/stream
MODEL_PATH=/Users/john/dev/ggml-small.en.bin

3. Start the application

yarn start

4. Start speaking!

You should have the application ready now and you can start using it as shown in the video below. Make sure to unmute 🔉 the video before playing.

voice-writing-demo.mp4

📦 Distributing as an app

1. Ensure .env has all the values as .env file gets packaged with the application.
2. Run yarn build. It builds the app only using target: dir on Mac. Use target: dmg to build dmg.
3. Check dist folder for the app.