#3838th of 38 in VoiceFor voice, speech and music

vosk-api

Offline speech recognition toolkit with streaming API and many language bindings

Stars
15.2k
License
Apache-2.0
Last commit
Oct 2026
Last release
Apr 2024
Language
Jupyter Notebook
Stars, last 29 weeks+840
31 Mar 2026: 14,350 stars7 Apr 2026: 14,396 stars14 Apr 2026: 14,436 stars21 Apr 2026: 14,488 stars28 Apr 2026: 14,530 stars5 May 2026: 14,561 stars12 May 2026: 14,596 stars19 May 2026: 14,639 stars26 May 2026: 14,685 stars2 Jun 2026: 14,717 stars9 Jun 2026: 14,744 stars16 Jun 2026: 14,767 stars23 Jun 2026: 14,789 stars30 Jun 2026: 14,811 stars7 Jul 2026: 14,839 stars14 Jul 2026: 14,869 stars21 Jul 2026: 14,908 stars28 Jul 2026: 14,949 stars4 Aug 2026: 14,967 stars11 Aug 2026: 14,991 stars18 Aug 2026: 15,019 stars25 Aug 2026: 15,034 stars1 Sept 2026: 15,060 stars8 Sept 2026: 15,084 stars15 Sept 2026: 15,102 stars22 Sept 2026: 15,116 stars29 Sept 2026: 15,136 stars6 Oct 2026: 15,147 stars13 Oct 2026: 15,166 stars20 Oct 2026: 15,190 stars

14,350 on 31 Mar 202615,190 on 20 Oct 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Vosk is an offline speech recognition toolkit covering 20+ languages and dialects. It exposes a streaming API with continuous large-vocabulary transcription, reconfigurable vocabulary and speaker identification, with bindings for Python, Java, Node.js, C#, C++, Rust, Go and others. Models are described as about 50 MB, and it runs from Raspberry Pi and Android up to clusters.

Who it is for: Developers adding offline speech-to-text to apps or devices

Strengths

  • Runs fully offline; no cloud service needed for transcription
  • Small models (about 50 MB) suit Raspberry Pi and Android
  • Bindings for Python, Java, Node.js, C#, C++, Rust, Go and more
  • Streaming API with reconfigurable vocabulary and speaker identification

Weaknesses

  • README gives no accuracy or word error rate figures
  • Last release was April 2024, despite recent commits
  • Install steps, examples and model downloads live on an external site
  • Speech-to-text only; no synthesis or diarization details in the README

What it needs

  • Models: Vosk offline speech recognition models (20+ languages)

Also in Voice

See all 38
Also in Voice
RankProjectScore
1voicestudioDesktop app for voice cloning, dubbing and dictation with local speech models79.5k starsAGPL-3.083 out of 100
2whisper.cppC/C++ port of OpenAI Whisper for local speech-to-text transcription54.5k starsMIT77 out of 100
3VoiceboxLocal voice studio for cloning, TTS, dictation and agent speech58.1k starsMIT71 out of 100
#4buzzOffline desktop app for transcribing and translating audio with Whisper22.1k starsMIT69 out of 100
#5meetilyDesktop app that records, transcribes and summarizes meetings locally Live demo ↗ (opens in a new tab)32k starsMIT68 out of 100