Intelligent voice dictation rooted in Arabic · Beta
Speak in Arabic dialects or English and watch your words transcribe in real time into precise, polished text across any desktop application — with 100% on-device privacy.
Windows, Ubuntu, Fedora & Universal Linux ready · Public Beta Edition
Your space to write
The same distraction-free workspace, whether you are dictating your first quick thought or refining a full document.
Speak naturally in your dialect. The app understands the real context of your words while preserving your authentic voice and tone.
Speak what you want from AI assistants and use smart drafting modes to transform rough speech into formatted instructions and actionable tasks.
Technical Specifications
Verified benchmarks and architectural guarantees for local Arabic speech processing.
| Feature | Nova (Beta Edition) | Standard Cloud Tools |
|---|---|---|
| Audio Privacy & Retention | 100% On-Device · Zero audio stored on intermediate servers | Audio stored on cloud servers and used for model training |
| Dialect Recognition | Modern Standard Arabic + Egyptian, Gulf, Levantine, North African | Limited dialect accuracy, frequent misrecognition |
| Automatic Diacritics & Grammar | Real-time punctuation, hamza correction, and diacritic marks | Raw unpunctuated text requiring manual editing |
| Response Latency | Sub-second instantaneous local transcription | Network round-trip latency and queue delays |
| App Integration | Universal global shortcut pastes directly into any active app | Isolated inside web browser or walled garden apps |
| Pricing & Licensing | Free Public Beta with transparent BYOK model | Expensive monthly subscription locks |
How It Works in 3 Steps
Lightweight, friction-free architecture. Speak freely in any window and let Nova handle the rest.
Enter your private connection key (Groq, OpenAI, or Gemini) once. Stored encrypted in your OS secure keyring with zero middleman.
Trigger Nova with a global hotkey inside any active app (Word, browser, email). Speak naturally in Arabic dialects or English.
Your speech is transcribed, punctuated, and grammatically corrected in sub-second time, pasting directly at your cursor.
Before you begin
Choose the workflow that fits your day, and understand how the application operates before your first recording.
Never. The app does not save or store your voice on intermediate servers. Audio streams directly and encrypted using your own private key.
Yes! Modern speech recognition easily understands everyday Arabic dialects, modern standard Arabic, and English with great fluidity.
You can insert text directly or copy it with a single click into Word, editors, email clients, and messaging apps.
You can generate a connection key in minutes from your preferred AI speech provider and paste it into the app settings.
Nova runs smoothly on everyday PCs running Windows 10/11 or Linux. A dedicated GPU is not required; the engine is optimized for standard CPUs with minimal memory overhead.
This is a beta release. Send a report to the Nova team or check documentation for help. Report a problem
Start your next writing
The application is free and lightweight. It connects securely using your private key on your device.
