The honest VoiceInk alternative

What this comparison is, and who built it

The boring truth is that VoiceInk is good. It is open-source under GPL-3.0, built in public on GitHub, and you can read every line, audit it, or run your own build. That is not a marketing line. It is a real feature, and it is one we do not offer.

If you value being able to inspect the code that listens to your microphone, VoiceInk has an argument that no closed-source app, including mine, can match. I am not going to wave that away to make my own product look better.

For everyone still reading, the gap is platform and the surrounding app. VoiceInk runs on Macs with Apple Silicon. We run on Windows too, and we wrap the whole thing in a managed account with history, presets, and an optional cloud layer. That is the entire argument, and I will spend the rest of the article showing it instead of asserting it.

What VoiceInk actually does

VoiceInk runs on macOS 14.4 or later, requires Apple Silicon, and is now also on iOS. It does live dictation. You hold a push-to-talk shortcut, speak, and the text lands at your cursor in whatever app you are in.

Transcription runs locally on the Apple Neural Engine, and the model lineup includes Parakeet v3, so your voice data never leaves your Mac. There is an optional Cloud Enhancement layer for people who want it, and it is genuinely optional. The local path is the default and it is fully private.

The part worth slowing down on is the license. VoiceInk is open-source under the GNU General Public License v3.0. The maintainer ships a prebuilt binary you can download from the site, and also publishes the source with build instructions, so you can either grab the ready-made app or compile it yourself. A lot of comparison posts get this wrong and claim you have to build it from source. You don't. The download is right there, which is more than I can say for half the open-source projects I have tried to install on a Sunday afternoon.

Pricing is one-time, with no subscription, across three lifetime tiers and a 14-day money-back guarantee. I am not quoting the figures here, because pricing pages move and you should read theirs and ours straight from the source. The shape of it is what matters. You pay once, you own it, and there is a refund window if it is not for you.

What you also get with Whisper by Remskill

Here is the part where I describe the thing I built, then let you judge the table. Whisper by Remskill is two products on one hotkey. The free tier is the whole local pipeline. You get the 8 Whisper transcription models, the Parakeet engine, fully-offline AI cleanup through Ollama, transcription history, presets, hotwords, hardware acceleration, model downloads, and a custom hotkey. No payment method at signup. You make an account, download the app, press the hotkey, and talk.

You pick your local engine based on what you need, not what we push. Whisper gives you 99 languages, translate-to-English, custom vocabulary, beam-size control, and hotword biasing, at the cost of speed. Parakeet is the NVIDIA TDT engine, about 600 MB, and it runs 5 to 10 times faster than Whisper on a CPU, covering English plus 24 other European languages. Pick Parakeet for speed and English. Pick Whisper for languages, translation, or fine control. Neither one is the default. That is your call, not ours.

Local accuracy typically lands between 95% and 99%, and it all runs on your CPU with no GPU required. The app itself is about 25 MB on disk, which is smaller than most of the screenshots people send me of it.

The Pro tier adds the Cloud surface, and it is bring-your-own-key. You paste your own OpenAI key, you pick the transcription model, and the AI enhancement runs on gpt-5-mini by default. There is also web search at your cursor through OpenAI's Responses API, which pastes a synthesised answer in a few seconds. We take no cut on top of OpenAI's rate. Your key, your bill.

The default hotkey is Ctrl+Space on Windows, remappable, with a push-to-talk chord on Mac. One account covers up to 3 devices.

Now the honest part, the trade we make against VoiceInk. We are closed-source. You cannot read our code or compile your own build, and if that is a dealbreaker for you, it is a real one. What we offer in return is Windows support and a managed account around the dictation. Different things matter to different people.

VoiceInk vs Whisper by Remskill, side by side

The table nobody else seems to fill with real rows. No prices in it — check both pricing pages for those.

A few honest reads of this table. VoiceInk is open-source and you can audit or rebuild it, which we cannot offer. It runs on the Neural Engine and ships an iOS app, while we have no mobile app at all. Those are not small wins. Everywhere the row is about Windows, a managed account, a stated language count, or a free-forever local tier, the gap is on our side of it.

Where the real difference lives, free local on either OS

This is what most people came here to compare, so let me be concrete. VoiceInk gives you private, local dictation on the Neural Engine. Your audio never leaves the Mac, and you pay a one-time price to keep it. There is a free trial first, then the license. If you are on Apple Silicon and that price is fine, it is a clean deal. Pay once, own it, audit the code if you want to.

Whisper by Remskill takes a different shape. The entire local pipeline is the free tier, with no expiry and no word cap, on Windows and on Apple Silicon Macs alike. Whisper transcription, Parakeet transcription, offline Ollama AI rewrites, history, presets, hotwords, model downloads. All of it, for as many words as your microphone can produce. You can dictate for eight hours on a flight in airplane mode, and the only ceiling is your laptop battery. The only thing the free tier leaves out is the Cloud surface — OpenAI transcription, gpt-5-mini enhancement, and web search — and that is the part you upgrade for.

Here is the one opinion I will spend in this article. Open source is a genuine feature, not a checkbox, and it is the right reason to pick VoiceInk. I am not going to wave it away to make my own product look better. Being able to read the code that hears your voice is worth something real. Our counter is not that open source doesn't matter. It is that most people want their dictation to run on Windows, with an account that remembers their settings across machines, and they are happy to trade the source for that. We give the local pipeline away for free on both operating systems and charge only for the optional cloud layer.

Most AI dictation apps are a Whisper API call, a tidy UI, and a recurring invoice. VoiceInk is not that. It is honest, local, and one-time. We are not that either. The difference between us and VoiceInk is which operating system you are on and whether you want to read the source, not who is quietly billing you for a wrapper.

When to stay on VoiceInk

This section earns the rest of the article. There are real reasons to stay, and I am not going to pretend otherwise.

For everyone else — Windows users, people who want a managed account with history across devices, anyone who wants a stated language count and translate-to-English, or who just wants to start free and never hit a wall — start with our free tier and see whether you ever need more. On the local pipeline, there is no wall. If you are on Windows specifically, the practical starting point is our walkthrough of voice to text on Windows, and if you are weighing another Mac option, the superwhisper alternative comparison covers that one.

If you only remember one thing

The thing voice typing fixes is the gap between having something to say and getting it into the document. VoiceInk closes that gap well on Apple Silicon, with code you can read, for a one-time price. We close it on Windows and Mac, give the local part away, and ask you to trust a closed app in return. Pick the one whose platform and model match how you actually work. If you are on Windows, or you just want to start free, the free one costs nothing to find out.