VoiceInk local models: the full catalog, importing your own Whisper file, and Ollama for cleanup

By Ryleigh Newman · Published October 11, 2026

VoiceInk can transcribe entirely on your Mac, and it gives you more local models to pick from than most dictation apps. You download them in AI Models, Local, then choose one per Mode. This page lists what is in the catalog, which models show text as you speak, how to import your own Whisper file, and how to keep the cleanup step local with Ollama.

I built YapToText, a dictation app for the Mac, so read what I say about VoiceInk with that in mind. Every fact about VoiceInk here comes from its own docs, read on October 11, 2026.

This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

Parakeet and Nemotron

From VoiceInk's docs, Local Models, read October 11, 2026. The Mode's language picker shows the exact choices.

Whisper, Cohere Transcribe, SenseVoice and Apple Speech

• Whisper. Tiny, Base, Large v2, Large v3, Large v3 Turbo and a quantized Large v3 Turbo, with English-only versions of Tiny and Base. Whisper processes the recording after you stop, with no partial text.
• Cohere Transcribe. On device, for Arabic, Chinese, Dutch, English, French, German, Greek, Italian, Japanese, Korean, Polish, Portuguese, Spanish and Vietnamese.
• SenseVoice Small. English, Japanese, Korean, Cantonese and Mandarin Chinese.
• Apple Speech. The Mac's own speech framework, on device. It needs macOS 26 or later and has no real-time option in VoiceInk. macOS manages its language downloads and limits how many languages you can hold.

The best VoiceInk model page helps choose between them, and Parakeet against Whisper explains the two main families.

Download, use and remove a model

1. Open AI Models, Local.
2. Download the model.
3. Open Modes, pick a Mode, and choose the model under Transcription, Model.
4. To free space later, use the model card's menu to delete the files or reveal them in Finder.

VoiceInk warns that local models do not work reliably on Intel Macs. Apple silicon is the recommended hardware. Larger models use more storage and take longer to load.

Import your own Whisper model

In AI Models, Local, choose Import Local Model and select a whisper.cpp-compatible .bin file. VoiceInk treats it as a Whisper model and marks it multilingual by default. The docs are clear that this is not a general importer for other model formats. Choose the imported model in a Mode afterward.

Keep the cleanup local too

Transcription is one step. Cleanup, which VoiceInk calls enhancement, is another, and a local transcription model does not make it local. The Local tab has three options for it:

• VoiceInk Refine. Built-in cleanup on a supported Mac. The Refine page covers it.
• Ollama. Connect to a model server, at http://localhost:11434 by default. A local server with a local model keeps enhancement on your Mac.
• Local CLI. Runs a command you installed with VoiceInk's prompt. The docs note such a command can still call a hosted service.

The VoiceInk privacy page lists what each choice sends.

Local models in YapToText

VoiceInk from its docs, read October 11, 2026. YapToText from its own settings.

VoiceInk offers far more model choice. YapToText offers one path that works without choosing. The local Whisper guide explains running Whisper on a Mac in general.

This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

The short version

• Where. AI Models, Local, then pick it in a Mode.
• Fast and local. Parakeet V3, the onboarding default.
• Live text. Parakeet optionally, Nemotron always.
• Your own model. Import a whisper.cpp .bin file.
• Local cleanup. Refine or Ollama.

If VoiceInk changes its model catalog, open an issue on GitHub and I will update this page. Every one gets read.

This app is completely free. You don't have to make an account. There are no subscriptions, no difficult downloads, and nothing is locked. Made for Apple silicon Macs. Source on GitHub.

About the author. Ryleigh Newman is a developer whose hands make typing difficult and who dictates every day with YapToText, the app this site describes. Facts about other apps come from their own sites, with the date they were read. More about the project.

Related

• VoiceInk guides: every page on this site about VoiceInk, by topic
• Which VoiceInk model to choose, from VoiceInk's own guidance
• VoiceInk cloud providers: which ones transcribe, how to add an API key, and who pays
• VoiceInk Refine and AI Enhancement: clean up dictation locally, and when you need a provider
• VoiceInk privacy: what stays on your Mac, what is sent, and the three settings to check
• How to use VoiceInk on a Mac: from the first dictation to Modes, the Dictionary and History
• Parakeet vs Whisper for dictation on a Mac: NVIDIA's model against OpenAI's
• Local Whisper on a Mac: what it needs, what it costs, and how to run it
• VoiceInk real-time transcription: which models stream, and how to turn it on

Sources

• https://tryvoiceink.com/docs/local-models (checked 2026-10-11)
• https://github.com/ryleighnewman/YapToText (checked 2026-10-11)
• https://apps.apple.com/us/app/yaptotext/id6786382289 (checked 2026-10-11)

Questions