## Full Description (4000 chars max)
Anti-Vocale transcribes voice messages from WhatsApp, Telegram, and other messaging apps directly on your Android device. All processing happens on-device — your audio never leaves your phone.
**AI MODELS**
Choose from multiple models:
- Whisper Turbo (recommended): Fast, accurate, 99 languages
- Whisper Medium: Very accurate, multilingual
- Whisper Small: Compact, 99 languages
- Distil Italian: Optimized for Italian, fastest
- Gemma 4 E2B: Multimodal AI, 140+ languages
- Gemma 4 E4B: Higher quality multimodal, 140+ languages
- Gemma 3n E2B: Multimodal AI, 140+ languages
- Gemma 3n E4B: Higher quality multimodal, 140+ languages
- Parakeet TDT: Lightweight, 25 European languages
- Qwen3-ASR 0.6B: Compact, 52 languages
**FEATURES**
- Share voice messages from any app
- Smart notifications with one-tap copy or share back
- Per-app notification settings (auto-copy, quick share)
- Transcription log with search, swipe actions, and undo
- Parallel chunk transcription and silence stripping for speed
- Customizable inference prompts for Gemma models
- Configurable inference threads with CPU auto-detection
- Resume interrupted downloads, HuggingFace auth for gated models
- Tasker integration for automation
**HOW IT WORKS**
1. Download an AI model (Whisper Turbo recommended)
2. Share a voice message from any messaging app
3. Get the transcription in a notification instantly
4. Copy or share it back to the conversation
**PRIVACY**
No network, no accounts, no data collection, no ads, no tracking.
**OPEN SOURCE**
Source code: https://github.com/RisorseArtificiali/anti-vocale
**REQUIREMENTS**
Android 8.0+ | 4GB+ RAM recommended | 350MB-4.5GB storage
**MODEL SIZES**
Whisper Turbo ~988MB | Whisper Medium ~903MB | Whisper Small ~358MB | Distil Italian ~939MB | Gemma 4 E2B 2.6GB | Gemma 4 E4B 3.7GB | Gemma 3n E2B 3.3GB | Gemma 3n E4B 4.2GB | Parakeet TDT ~464MB | Qwen3-ASR ~938MB
Transcribe voice messages offline. No internet required.