Features
Everything you need to stop typing
Voxa combines advanced speech recognition, Arabic dialect intelligence, and privacy-first architecture into a tool that works everywhere you do.
AI-Powered Dictation
Speak naturally and let Voxa transcribe your words with exceptional accuracy. Powered by whisper.cpp with Metal GPU acceleration, delivering results in under 700ms.
22 Arabic Dialects
The first dictation tool built specifically for Arabic. Egyptian, Gulf, Levantine, Moroccan, Iraqi, Sudanese, Yemeni, and 15 more dialects — each treated as its own language model.
Filler Word Removal
Voxa automatically strips out ums, uhs, likes, and repetitions. Your speech becomes clean, professional text without manual editing.
Custom Dictionary
Voxa learns the names, technical terms, slang, and dialect-specific expressions you use. Your dictionary gets smarter with every dictation session.
100+ Languages
From Arabic to Mandarin, Spanish to Hindi. Voxa handles code-switching seamlessly — mix languages mid-sentence and get accurate transcription with proper BiDi text handling.
Fully Offline
Every word is processed on your device. No internet required. Works on planes, in tunnels, in remote locations. Your voice data never touches a server.
Works in Any App
One hotkey. Any app. Gmail, Slack, VS Code, WhatsApp, Notion, Word — press, speak, release. Clean text appears wherever your cursor is.
Local AI Processing
Privacy is our architecture, not a feature. All speech recognition runs on-device using your GPU. We have no servers to receive your audio. We physically cannot listen.
Free to start. No credit card required.