VoxBoard on-device models
Mirror of the model archives that the VoxBoard iOS app downloads for fully on-device dictation and text cleanup. Files are unchanged repackagings of the upstream releases below; each archive's SHA-256 is pinned in the app.
Every model keeps its original license. Using a file means accepting the
license of its upstream model. See LICENSE.md.
| Folder / file | Upstream model | License |
|---|---|---|
whisperkit/openai_whisper-*.zip |
OpenAI Whisper, Core ML conversion by Argmax (argmaxinc/whisperkit-coreml) | MIT |
parakeet/parakeet-tdt-0.6b-v2-coreml.zip, -v3-coreml.zip |
NVIDIA Parakeet TDT 0.6B, Core ML conversion by FluidInference (FluidInference) | CC BY 4.0 |
moonshine/* |
Moonshine by Useful Sensors / Moonshine AI (moonshine-ai) | See upstream repository |
llm/qwen3-*, llm/qwen35-* |
Qwen3 / Qwen3.5 by Alibaba Qwen (Qwen), GGUF quantization | Apache 2.0 |
llm/gemma3-* |
Gemma 3 by Google (google), GGUF quantization | Gemma Terms of Use: https://ai.google.dev/gemma/terms |
llm/llama32-* |
Llama 3.2 1B by Meta (meta-llama), GGUF quantization | Llama 3.2 Community License: https://www.llama.com/llama3_2/license/ — Built with Llama |
llm/smollm2-* |
SmolLM2 by Hugging Face (HuggingFaceTB), GGUF quantization | Apache 2.0 |
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support