Live Audio Transcription

Real-time audio transcription with Go backend and web frontend. Captures microphone audio, streams via WebSocket, and provides live transcription using Google Cloud Speech-to-Text and Vertex AI.

Quick Start

Prerequisites

• Go 1.24+ installed
• Google Cloud account with an active project
• Google Cloud CLI (gcloud) installed

1. Google Cloud Setup

Install and configure gcloud CLI:

# Install gcloud CLI (if not already installed)
# Visit: https://cloud.google.com/sdk/docs/install

# Initialize gcloud and authenticate
gcloud init

# Set up Application Default Credentials
gcloud auth application-default login

Create/configure your GCP project:

# Set your project ID (replace with your actual project)
export GCP_PROJECT_ID=your-project-id
export GCP_LOCATION=us-central1

# Set the project as default
gcloud config set project $GCP_PROJECT_ID

# Enable required APIs
gcloud services enable speech.googleapis.com
gcloud services enable aiplatform.googleapis.com

Set environment variables:

# Add these to your shell profile (.bashrc, .zshrc, etc.)
export GCP_PROJECT_ID=your-project-id
export GCP_LOCATION=us-central1

# AI Model Configuration
export GEMINI_MODEL=gemini-2.5-flash  # Gemini model to use (default: gemini-2.5-flash)

# Logging Configuration
export LOG_LEVEL=INFO    # DEBUG, INFO, WARN, ERROR (default: INFO)
export LOG_FORMAT=JSON   # JSON, TEXT (default: JSON)

# Preset Configuration
export PRESET_DIRECTORY=./presets  # Directory containing preset files (default: ./presets)

# Optional: Set custom port (default: 8080)
export PORT=8080

2. Run the Application

# Install Go dependencies
go mod tidy

# Run the server
go run main.go

3. Access the Web Interface

Open your browser and navigate to: http://localhost:8080

HTTPS Support (Optional)

For secure connections, the server automatically detects SSL certificate files and enables HTTPS:

Note: secure connection is mandatory to use anything else that localhost

Generate Self-Signed Certificate

# Generate a private key
openssl genrsa -out server.key 2048

# Generate a self-signed certificate (valid for 365 days)
openssl req -new -x509 -key server.key -out server.crt -days 365

# You'll be prompted for certificate details:
# - Country Name: US
# - State: Your State  
# - City: Your City
# - Organization: Your Organization
# - Organizational Unit: IT Department
# - Common Name: localhost (IMPORTANT: use 'localhost' for local development)
# - Email: your-email@domain.com

Note: When prompted for "Common Name", enter localhost to avoid browser certificate warnings during local development.

HTTPS Access

Once certificate files (server.crt and server.key) are present in the project directory:

• Server automatically starts in HTTPS mode
• Access via: https://localhost:8080
• WebSocket connections use: wss://localhost:8080/ws

Certificate Security

⚠️ Self-signed certificates will show browser warnings. For production use, obtain certificates from a trusted Certificate Authority (CA) like Let's Encrypt.

Features

• Real-time speech transcription with interim/final results
• Multi-language support with auto-detection
• AI-powered summarization with configurable Gemini models
• Audio visualization and session statistics
• Copy transcripts and summaries to clipboard
• Pre-configured prompt presets for different use cases (meetings, interviews, lectures)
• Configurable logging with structured output

Configuration

• Languages: Configure BCP-47 codes (default: en-US,fr-FR,es-ES)
• Logging: Set LOG_LEVEL (DEBUG/INFO/WARN/ERROR) and LOG_FORMAT (JSON/TEXT)
• Port: Set PORT environment variable (default: 8080)
• Audio: 16kHz LINEAR16 mono format

Presets

The application supports prompt presets for different use cases. Presets are stored as text files in the presets directory (configurable via PRESET_DIRECTORY environment variable).

Built-in Presets

• General Summary: Basic conversation summarization
• Meeting Summary: Business meeting focused with decisions and action items
• Interview Summary: Job interview evaluation and assessment
• Lecture Notes: Educational content with key concepts and takeaways

Custom Presets

Create custom presets by:

1. Setting PRESET_DIRECTORY environment variable to your custom directory
2. Creating .txt files following this format:

Title: Your Preset Title
Summary: Your summary prompt content here...
This can span multiple lines.

Conclusion: Your conclusion prompt content here...
This can also span multiple lines.

3. Restarting the application to load new presets

API Endpoints

• GET / - Web interface
• GET /api/default-prompt - Returns the default summary prompt as JSON
• GET /api/presets - Returns available preset names and titles as JSON
• GET /api/presets/{name} - Returns specific preset content (title, summary, conclusion)
• WebSocket /ws - Real-time audio streaming and transcription

Build

go build -o live_transcription main.go
./live_transcription