Voice2Text AI

A cross-platform Python application that transcribes voice input using GPU-accelerated Whisper, sends text to local Ollama AI models for intelligent responses, and reads the output aloud with text-to-speech.

Features

• 🎙️ Real-time voice transcription with OpenAI Whisper (GPU-accelerated)
• 🤖 AI chat integration with local Ollama models
• 🔊 Text-to-speech output with pause/resume
• 🎨 Modern dark gradient UI
• 📦 Cross-platform executables (Windows/macOS/Linux)
• 🖥️ Flatpak support
• ⚡ Optimized performance with silence detection
• 🔄 Automatic retry logic for network requests

Requirements

• Python 3.8+ (maybe, I think the standalone app runs without it)
• Ollama running locally (http://localhost:11434) (only for AI interaction)
• Microphone access (The whole point is dictation)
• CUDA-compatible GPU (optional, for faster Whisper processing. Highly recommended.)

Installation

Pre-built Executables

Download the latest release from GitHub Releases:

• Windows: Voice2Text.exe Untested
• macOS: Voice2Text Untested
• Linux: 'Voice2Text' https://drive.google.com/file/d/1MmF6Vr_3nz1yket2SdnHCI_Od5OpIzQd/view?usp=sharing

Linux (Flatpak) (ToDo)

# Install from Flathub (when available)
flatpak install flathub com.voice2text.app

# Or download from GitHub Releases
# Download: voice2text.flatpak from latest release
# Install: flatpak install --user voice2text.flatpak
# Run: flatpak run com.voice2text.app

From Source (This might actually work)

git clone https://github.com/crhy/Voice2Text-AI.git
cd Voice2Text-AI
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Development

# Run tests
python test_voice.py
python test_mic.py
python test_whisper.py

# Build executable
pyinstaller voice_app.spec

# Build Flatpak (This has never worked.  Yet)
flatpak-builder --force-clean build com.voice2text.app.yml
flatpak-builder --user --install build com.voice2text.app.yml

# Export Flatpak for distribution
flatpak build-bundle build voice2text.flatpak com.voice2text.app

Setup Ollama

# Install Ollama (if not already installed)
curl -fsSL https://ollama.ai/install.sh | sh

# Start Ollama service
ollama serve

# Pull a model (in another terminal)
ollama pull llama3.2  # or any preferred model

Usage

Running the App

• Executable: Right Click, Select "Allow Executing File as Program", Double-click the downloaded file (wait, it's slow to load. Once it's up it's crazy fast)
• From Source: python voice_app.py
• Flatpak: flatpak run com.voice2text.app (Good luck?)

Quick Start

1. Select your Ollama model from the dropdown
2. Click "🎙️ Start Dictation"
3. Speak clearly into your microphone
4. Click "⏹️ Stop Dictation" when finished
5. Click "🤖 Query AI" to get AI responses
6. Responses are displayed and spoken aloud

The app automatically saves your settings and provides real-time status updates.

Notes

• Whisper models run locally (internet required for initial download)
• Edge TTS provides local speech synthesis
• GPU acceleration speeds up transcription significantly
• Settings persist between sessions
• App works offline after initial model downloads

Troubleshooting

• Microphone: Run python test_mic.py
• Ollama: Ensure it's running at http://localhost:11434
• GPU: Check with nvidia-smi (optional)
• Audio: Grant microphone permissions
• Models: First run downloads Whisper models (~2GB)

Project Structure

Voice2Text-AI/
├── voice_app.py          # Main GUI application
├── main.py              # Alternative Tkinter version
├── voice_app_kivy.py    # Kivy mobile version
├── voice_to_opencode.py  # CLI version
├── requirements.txt      # Python dependencies
├── test_*.py            # Test scripts
├── *.spec               # PyInstaller configs
├── *.desktop            # Linux desktop files
├── com.voice2text.app.*  # Flatpak manifests
├── config.json          # App configuration
├── voice_config.json    # Voice app settings
├── logo.png             # Application logo
├── install.sh           # Linux installer
└── README.md

Distribution

Building Releases

# Create GitHub release with all platform binaries
# Upload these files to GitHub Releases:
# - Voice2Text.exe (Windows)
# - Voice2Text (macOS)
# - voice2text.flatpak (Linux)

Flathub Submission

To submit to Flathub for official distribution:

# Fork the Flathub repository
# Add your manifest to: https://github.com/flathub/flathub
# Submit pull request with com.voice2text.app.yml

Contributing

Contributions welcome! Please submit issues and pull requests on GitHub.

License

This project is open source. See individual files for license information.