A real-time speech-to-text (STT) transcription project written in Go, leveraging Deepgram API for accurate, low-latency audio transcription.
Perfect for audio streaming, voice notes, or voice-controlled applications.

• 🎧 Real-time audio streaming from microphone
• 📝 Live transcription using Deepgram API
• 🌐 Supports multiple languages (en, es, fr, etc.)
• ✅ Graceful shutdown on Ctrl+C
• ⚡ Uses PortAudio for cross-platform audio capture
• 📦 Easy to configure via .env

# Run with default English language
go run main.go --lang=en

# Run with Spanish
go run main.go --lang=es

• Prerequisites

• Go 1.20+

• PortAudio (install via brew, apt, or other package manager)

• Deepgram API Key

# MacOS
brew install portaudio

# Ubuntu
sudo apt-get install portaudio19-dev

git clone https://github.com/Kesha005/audio-stt.git
cd audio-stt
go mod tidy
cp .env-example .env
# Set DEEPGRAM_KEY in .env

go run main.go --lang=en
go run main.go --lang=es

Use languages which support deepgram stt. Deepgram stt https://developers.deepgram.com/docs/models-languages-overview.

2025/11/21 15:28:24 Opened
2025/11/21 15:28:29 Received transcription: Salome. How are you?
2025/11/21 15:28:39 Received transcription: Hi. How are you?

Deepgram for the STT API

PortAudio for cross-platform audio streaming

Go community for amazing tools and libraries