TECHSY.communityby techsy.io
Sign Up
All open source
Open SourceVoice & Media11k

Vosk

Lightweight offline speech recognition toolkit for 20+ languages and embedded devices.

Vosk is an offline speech recognition toolkit with bindings for Python, Java, Node, C#, and more, and models for over twenty languages and dialects. Models are small (around 50 MB for the compact variants), making it well suited for mobile, Raspberry Pi, and on-device assistants while still supporting large server-grade models.

View repo

Repository

alphacep/vosk-api

Language

Jupyter Notebook

What you'd build with it

  • Embed offline voice command or dictation into a mobile or IoT product without cloud calls
  • Stream microphone audio to a local recognizer with low latency for a voice assistant
  • Run speaker-independent transcription in a language not well served by larger models

Tags

sttofflineembeddedmultilingual