Anti-Vocale APK

Anti-Vocale

Private, On‑Device Voice‑Message Transcription

• Anti‑Vocale turns voice notes from WhatsApp, Telegram, and any other Android messenger into readable text—without ever leaving your phone.

• All transcription runs locally, so your audio stays private.

Built‑In Speech Models

• Parakeet TDT (recommended) – Fast, supports 25 European languages. Choose High‑Quality (862 MB) or Compact (464 MB).

• Whisper Turbo – Balanced speed & quality for 99 languages.

• Whisper Distil Italian – Optimized for the best Italian transcription, tuned for short voice messages.

• Whisper Medium / Small – Higher accuracy or a smaller download, respectively.

• Nemotron 3.5 Streaming – Real‑time text appears as it decodes; 45 languages.

• Qwen3‑ASR – Tiny footprint, covers 59 languages.

• GigaAM v3 – Ultra‑small, ultra‑fast Russian model.

• Gemma 4 / 3n (LiteRT) – Multimodal AI with customizable prompts.

Bring Your Own Models

• Import any ONNX‑compatible transcription model from local storage or a URL (Whisper fine‑tunes, CTC, SenseVoice, Sherpa‑ONNX, etc.).

• Automatic integrity checks, support for split‑file models, and dedicated share targets for each import.

• Perfect for languages not covered by the bundled models.

Core Features

• One‑Tap Share – Send any voice message to Anti‑Vocale via the Android share sheet.

• Smart Notifications – Copy, re‑share, or scroll through long transcripts without opening the app.

• Auto‑Copy & Auto‑Save – Transcripts are copied to the clipboard and saved to a folder of your choice (Drive, Syncthing, …).

• Video Transcription – Extracts the audio track and creates a text version.

• Streaming Text – See progressive results instantly with streaming models.

• Per‑App Notification Settings – Fine‑tune alerts for each messenger.

• Searchable Log – Swipe actions and full‑text search for past transcriptions.

• Speed Optimizations – Silence stripping and parallel chunk processing.

• Neural Acceleration – Uses NNAPI when available, with automatic CPU fallback.

• Tasker Integration – Automate transcription workflows.

Privacy

• No network connection required after the initial download.

• No accounts, ads, analytics, or tracking.

• Your audio never leaves the device.

Open Source

• Licensed under Apache 2.0.

• Source code: https://github.com/RisorseArtificiali/anti-vocale

• System Requirements

• Android 8.0 or newer

• Recommended: 4 GB RAM

• Storage: ≈ 326 MB – 1 GB per model (depending on size)

FAQ

Q: Do I need an internet connection to transcribe my voice notes?

A: No. After you download the chosen speech model the first time, all processing happens offline. The app never sends your audio to a server, so you can transcribe even when you’re completely offline.

Q: Which speech models are bundled with Anti‑Vocale and how do I pick one?

A: The app ships with several ready‑to‑use models: Parakeet TDT (high‑quality or compact versions), Whisper Turbo, Whisper Distil Italian, Whisper Medium/Small, Nemotron 3.5 Streaming, Qwen3‑ASR, GigaAM v3, and Gemma 4/3n (LiteRT). You choose based on language coverage, speed, accuracy, and storage preferences—Parakeet TDT is great for European languages, Whisper Turbo covers 99 languages, and the distilled Italian model is tuned for short Italian voice notes.

Q: Can I use my own transcription model with Anti‑Vocale?

A: Absolutely. Anti‑Vocale supports importing any ONNX‑compatible model from local storage or a URL, whether it’s a Whisper fine‑tune, a CTC model, SenseVoice, Sherpa‑ONNX, or something else. The app verifies the model’s integrity and lets you set it as the active engine for transcription.

Version History

v1.11.3——10 Sep 2026

Transcribe voice messages offline. No internet required. Download the latest version of Anti-Vocale to enjoy new features and updates immediately!

*1.11.3

Minor bug fixes and improvements. Install or update to the newest version to check it out!

*1.11.2

Minor bug fixes and improvements. Install or update to the newest version to check it out!

*1.11.1

Minor bug fixes and improvements. Install or update to the newest version to check it out!