Cyberax AI Playbook
cyberax.com
← All models
Model · Alpha Cephei · Speech-To-Text

Vosk

An offline, lightweight speech recognition toolkit. Runs on phones, Raspberry Pi, and embedded devices — the right choice when Whisper is too heavy.

Modality
Speech-To-Text
License
Apache 2.0 (Open)
Released
January 1, 2020
Last verified
September 8, 2026
Runs locally
Yes

Strengths

  • Runs on truly constrained hardware (mobile, Raspberry Pi)
  • Streaming/realtime capable
  • Apache 2.0 with mature SDK bindings

Weaknesses

  • Accuracy below Whisper on hard audio
  • Languages are downloaded as separate models — bigger languages mean bigger downloads

Try it

WhereTypeNotes
Vosk models weights Apache 2.0 — many sizes available
GitHub local Bindings for Python, Java, Node, Swift, etc.

Official sources

Change log

  • — Verified actively maintained — real feature commits through 2026-08-09, despite no tagged release since v0.3.50 (2024-04-22). The release gap is not staleness. No successor project exists.
  • — Initial entry.

Related models

GPT Live Transcribe

OpenAI · Speech-To-Text

OpenAI's realtime transcription model — the streaming half of the pair replacing GPT-4o Transcribe. Built for live captioning and voice agents, with tunable latency and keyword hints.

Proprietary
Try on: OpenAI API
Released Jul 2026 · Verified 2026-09-08

GPT Transcribe

OpenAI · Speech-To-Text

OpenAI's current hosted transcription model and the named replacement for GPT-4o Transcribe. Priced per minute of audio rather than per token, for batch and file-based transcription.

Proprietary
Try on: OpenAI API
Released Jul 2026 · Verified 2026-09-08

NVIDIA Parakeet

NVIDIA · Speech-To-Text

NVIDIA's STT family. parakeet-tdt-0.6b-v2 tops the HF Open ASR leaderboard for English; parakeet-tdt-0.6b-v3 adds 25-language multilingual support. Very fast on NVIDIA hardware via NeMo.

CC BY 4.0 Open weights Runs locally
Try on: Hugging Face (English) · Hugging Face (multilingual)
Released May 2025 · Verified 2026-09-08

Whisper large-v3 Turbo

OpenAI · Speech-To-Text

OpenAI's distilled Whisper variant. ~8× faster than large-v3 with most of the accuracy retained — the practical default for high-throughput STT pipelines.

MIT Open Runs locally
Try on: Hugging Face · Replicate · Groq
Released Oct 2024 · Verified 2026-06-08

Whisper large-v3

OpenAI · Speech-To-Text

High-accuracy multilingual speech-to-text. Best-in-class for non-English audio; the de-facto open baseline.

MIT Open Runs locally
Try on: Hugging Face · Replicate · Groq
Released Nov 2023 · Verified 2026-06-08

Distil-Whisper

Hugging Face · Speech-To-Text

Hugging Face's distillation of Whisper (distil-large-v3.5). Faster than Whisper-large-v3-Turbo on long-form audio at small accuracy cost; English-only — pick when you don't need multilingual.

MIT Open Runs locally
Try on: Hugging Face
Released Nov 2023 · Verified 2026-09-08