Anti-Vocale

Anti-Vocale transcribes voice messages from popular messaging apps directly on your Android device. Share a voice note from your chat app and Anti-Vocale turns it into text on-device, so your audio never leaves your phone.

Note: Anti-Vocale is built around voice messages shared from messaging apps. Audio shared from other sources (for example a voice recorder) may use formats the app can't read yet.

No network connection, no cloud, no accounts, no tracking. AI models run locally via ONNX Runtime.

CHOOSE YOUR MODEL

Multiple swappable ASR backends, each with different tradeoffs:
- Parakeet TDT: Fast multilingual, 25 European languages
- Whisper Distil Italian: Best Italian quality, optimized for voice messages
- Whisper Turbo: 99 languages, balanced speed/quality
- Whisper Small: Compact, 99 languages
- Qwen3-ASR: Compact, 52 languages
- Nemotron 3.5 streaming: Progressive text display in real time
- Gemma 4 E2B/E4B: Multimodal AI, 140+ languages

FEATURES

- Share voice messages from any app via the Android share sheet
- Auto-copy transcription to clipboard (paste straight back into chat)
- Read long transcripts page-by-page in the notification (back/forward paging, no app needed)
- Auto-save transcripts to a folder of your choice (Drive, Syncthing, Dropbox)
- Video file transcription (extracts the audio track)
- Progressive display — text appears in real time as it decodes
- Per-app notification preferences (auto-copy, quick share back)
- Transcription log with search and swipe actions
- Silence stripping and parallel chunk processing for speed
- Tasker integration for automation

PRIVACY

No network, no accounts, no data collection, no ads. The audio never leaves your phone. Models download once from HuggingFace; pick one and you're set.

OPEN SOURCE

Apache 2.0 license. Source code: https://github.com/RisorseArtificiali/anti-vocale

REQUIREMENTS

Android 8.0+ | 4GB+ RAM recommended | 350MB-940MB storage per model