Directory › Speech (STT / TTS) › Vosk

Vosk

Speech (STT / TTS) ✓ verified 🔒 local-only Apache-2.0
GitHub ↗

Vosk is an offline speech recognition toolkit that uses small models for many languages. It can be used to recognize spoken words and phrases from audio recordings. The tool is self-hostable.

Who it's for: This tool is suitable for individuals or organizations looking for a simple, offline speech recognition solution.

Strengths

  • Supports multiple languages
  • Small model sizes
  • Open-source

Trade-offs

  • Limited feature set compared to online services
  • May not perform well with noisy audio

Data locality

Runs fully offline — no data leaves your machine.