← VoiZ

VoiZ Open-source Credits

Updated November 13, 2026

VoiZ’s on-device dictation is built on open work by other people and companies. Thank you to everyone below. The same credits are in the app under Settings, Open-source credits.

Parakeet TDT 0.6B v3 (speech model)

© NVIDIA Corporation
License: CC BY 4.0
Source: huggingface.co/nvidia/parakeet-tdt-0.6b-v3

VoiZ for Mac on Apple-silicon uses an ONNX conversion of this model, compressed to 8-bit and published by a community contributor (istupakov). Changes made: the model was converted to a different file format and quantized so it runs on-device. NVIDIA does not endorse VoiZ.

Parakeet TDT 0.6B v3, ONNX version (conversion)

Converted by istupakov from the NVIDIA model above
License: CC BY 4.0
Source: huggingface.co/istupakov/parakeet-tdt-0.6b-v3-onnx

Whisper (speech models)

© OpenAI, converted for whisper.cpp by Georgi Gerganov
License: MIT License
Source: huggingface.co/ggerganov/whisper.cpp

Used on Intel Macs and as a fallback.

transcribe-rs (runs the speech models)

© 2025 Ilya Stupakov
License: MIT License
Source: github.com/cjpais/transcribe-rs

ort (ONNX Runtime for Rust)

© 2023–2026 pyke.io, © 2020 Nicolas Bigaouette
License: MIT License or Apache License 2.0
Source: github.com/pykeio/ort

whisper-rs (Whisper for Rust)

tazz4843 and contributors
License: Unlicense (public domain)
Source: codeberg.org/tazz4843/whisper-rs

Full license texts

The links above go to each license. If you would like a copy of any license notice, write to jorivan75@mkrlab.io.