Real-time English captions powered by offline April-ASR, with optional instant translation via Google Translate or DeepL. Designed for seamless live speech-to-text and multilingual translation, even without internet connectivity.

中文 vosk-android-demo

Offline free automatic speech recognition

An AI-powered teaching assistant that transcribes lecture videos offline using Vosk, generates semantic embeddings, and answers questions contextually using RAG with Groq or Gemini models.

Offline Voice Trigger (Hotword) system — powered by Faster-Whisper Low-latency speech command recognition that runs fully offline on Raspberry Pi, RK3566/3588, and x86 platforms. Uses WebRTC VAD for real-time speech detection, Faster-Whisper for transcription, and fuzzy matching to trigger local system commands.