TL;DR: If you want dictation whose source you can read, VoiceInk is the best finished app and whisper.cpp is the best engine to build on. Neither can type a prompt, press Enter, and switch apps for you the way Infina can on the three spoken words "type", "send it" or "enter", and "open", and Infina is not open source. This page is honest about that split: open the source or close the loop, and here is exactly what each choice costs.
Two different things get called open source
Most confusion in this search comes from mixing up the model, the engine, and the app.
- The model is the trained weights: Whisper, Parakeet, Vosk's models. Several are openly licensed.
- The engine is the code that runs the model on your hardware: whisper.cpp, Vosk's runtime.
- The app is the thing with a menu bar icon that puts text into your editor.
An app can be closed source while running fully open models on your machine. That is the most common arrangement on the Mac today, and it is also what Infina does. Being clear about which layer you actually need saves a lot of wasted evenings.
| Tool | Layer | License | Ready to dictate with |
|---|---|---|---|
| VoiceInk | App | GPLv3 | Yes, buy or build it |
| whisper.cpp | Engine | MIT | No, you build on it |
| OpenAI Whisper | Model and reference code | MIT | No, it is a research repo |
| Vosk | Engine and models | Apache 2.0 | No, for developers |
| Moonshine | Model | MIT | No, for developers |
| Talon | App | Closed core, open community scripts | Yes, after setup |
1. VoiceInk: the best open source dictation app you can actually use
VoiceInk is a macOS dictation app licensed under GPLv3 that transcribes locally with Whisper models. It is the only entry on this list that is both genuinely open and genuinely a finished app: a menu bar icon, a hotkey, text in your editor.
It sells for $25 to $49 once depending on how many Macs, with a 14-day money-back guarantee, and the source is public if you would rather audit it or build it yourself. Paying for software you could compile is a perfectly rational way to support a single developer, and that is what this is.
The limits are the limits of an indie project: hotkey-only, Apple Silicon on a recent macOS, and less polish than the commercial apps. It types text and stops there.
Pick it if: you want a working, auditable, local dictation app for the price of a lunch. Skip it if: you need hands-free operation. Compared directly in VoiceInk vs Infina.
2. whisper.cpp: the best engine to build on
whisper.cpp is a plain C and C++ port of Whisper inference, MIT licensed, with Apple Silicon acceleration. A large share of the "local Whisper" Mac apps in existence are wrappers around it, which tells you how good it is.
It is not a dictation app. There is no hotkey, no text insertion, no menu bar. What you get is a fast, dependency-light way to turn audio into text in your own program.
Pick it if: you are building something. Skip it if: you wanted to dictate this afternoon.
3. OpenAI Whisper: the model everything else stands on
The original Whisper repository is MIT licensed and includes the model weights that made local dictation practical on consumer hardware. It is a research release: Python, a command line, and no interface.
Run it once to understand what your dictation app is doing under the hood, then use a wrapper for daily work. More on which apps use it in Whisper dictation on Mac.
Pick it if: you want the reference implementation. Skip it if: you want a product.
4. Vosk: the best choice for small, always-listening tasks
Vosk is Apache 2.0 licensed, ships small models in many languages, and is built for streaming recognition with a tiny footprint. That makes it the usual answer when something needs to listen continuously without eating the machine, for example spotting a single trigger word.
Its accuracy on long free-form dictation is not where the big Whisper-class models are, and it is a library rather than an app.
Pick it if: you need lightweight, streaming, always-on recognition in your own code. Skip it if: you want a transcript of a meeting.
5. Moonshine: the newest small model worth knowing
Moonshine is an MIT-licensed family of small speech-to-text models aimed at fast recognition on modest hardware. It is young and firmly in developer territory, but it is the most interesting recent entry in the small-model space and worth watching if you are building.
Pick it if: you are prototyping on-device speech. Skip it if: you need something proven today.
6. Talon: powerful, free, but not open in the way people assume
Talon is free to download and its community command sets are public and extensively shared, which is why it reads as open source in a lot of forum posts. The core application itself is not. That distinction matters if your requirement is auditability rather than cost.
What it is, unambiguously, is the most capable voice control tool on the Mac for people who cannot use a keyboard, and it is free. See Talon vs Infina.
Where Infina fits, and where it does not
Infina is not open source, and we are not going to pretend otherwise on a page about open source.
What is true is that Infina runs open models on your own Mac: we did not train Parakeet or Vosk, we run them locally, and nothing about your dictation needs a server. So on the axis most people in this search actually care about, which is "does my audio leave the machine", Infina behaves the way the open tools do.
What you get for the closed source and the $99/year is the part none of the open tools do: three trigger words that close the loop. Say "type" and the prompt is typed. Say "send it", or a plain "enter", and it is submitted. Say "open" and you are in the next app. No key held down, no Return pressed, no hands. VoiceInk, whisper.cpp and the rest all stop at putting text on the screen.
There is also an optional cloud add-on at $5/month billed annually at $60 on top of the yearly plan for LLM-polished output; it is a post-purchase option and it is the only part of Infina that is not local.
When an open tool is the right call
- Auditability is a hard requirement. Then it is VoiceInk, full stop. No closed app can satisfy that, including ours.
- You are building a product. whisper.cpp or Vosk, not any app on this page.
- Your budget is zero and you will invest time. Talon, or Apple's built-in dictation.
- You want to modify how transcription behaves. Only the open stack lets you.
If what you want is the source, take VoiceInk and be happy. If what you want is to stop touching the keyboard, that is a different purchase.
FAQ
What is the best open source dictation app for Mac? VoiceInk. It is GPLv3, transcribes locally with Whisper models, and is a finished app rather than a library. It costs $25 to $49 once, and the source is public if you would rather compile it.
Is Whisper free to use? Yes. OpenAI released Whisper under the MIT license, including weights, which is why so many local Mac dictation apps are built on it. whisper.cpp is the MIT-licensed C and C++ implementation most of them actually ship.
Is Infina open source? No. Infina runs open speech models locally on your Mac, so your audio stays on the machine, but the app itself is closed and costs $99/year. If auditable source is your requirement, use VoiceInk.
Is Talon open source? Not the core application. Talon is free to download and its community command sets are public, which is where the confusion comes from, but the app itself is closed.
Can an open source tool type a prompt and press Enter for me? Not out of the box. VoiceInk and the Whisper-based apps insert text and stop. Talon can be scripted to press keys, and it is free, but you configure the commands. Infina does it on three spoken words with nothing to set up.
The bottom line
Working, auditable, local dictation today: VoiceInk. An engine to build on: whisper.cpp, with Vosk for small streaming jobs and Moonshine if you like living early. The reference: OpenAI's Whisper repo. Free and programmable, if not open: Talon.
If the real goal was privacy rather than source code, every tool here and Infina all keep audio on your Mac; the comparison then comes down to the loop, which is covered in on-device dictation on Mac and private dictation apps.
The short answer, by what you are trying to do
If you want to skip the detail above, find your sentence in this table and buy that one.
The tools this page covers (VoiceInk, Talon, Apple Dictation) are at the top. The rest of the field follows, because the right answer is sometimes none of the above.
| Tool | Price | Keyboard needed | Pick it if |
|---|---|---|---|
| Infina | $99 a year, unlimited words and unlimited commands | No. Speak "type", speak the prompt, say "send it" or "enter". | you want to prompt AI tools without touching the keyboard. Say "type" and the words go in, "send it" and the prompt goes, "open Cursor" and you are in the next agent. It is the only one of these that finishes the whole loop by voice in the tools developers prompt in all day. |
| VoiceInk | $25 to $49, paid once | Yes. A key to start dictating, Enter to send. | you want plain local dictation at the smallest possible price and you never need your voice to control the machine. |
| Talon | free | No, once you have learned its command language. | you want to drive the entire computer by voice, cursor and clicks included, and you are willing to learn a command language to get there. Talon is the serious hands-free tool on this list. Infina goes after the same freedom with three ordinary English words instead of a syntax to learn. |
| Apple Dictation | free, built in | Yes. A key to start dictating, Enter to send. | you dictate a sentence here and there and do not want to install anything. |
| Wispr Flow | $15 a month, or $144 a year billed annually | Yes. A key to start dictating, Enter to send. | you need dictation on an iPhone or an Android phone as well as a laptop, or you write in one of 100+ languages. Those are the two things Infina does not ship. |
| Superwhisper | $8.49 a month, $84.99 a year, or $249.99 once | Yes. A key to start dictating, Enter to send. | you want to pick and swap the speech model yourself, and dictation into a text box is the whole job. It stops where the text box ends: there is no spoken send, and no spoken app switch. |
| MacWhisper | 65 euros, paid once | Yes. A key to start dictating, Enter to send. | the job is transcribing recorded audio files rather than talking to your computer while you work. |
The part that is actually different: zero keypresses
The mainstream dictation apps (Wispr Flow, Superwhisper, MacWhisper, VoiceInk, Apple Dictation) are all push-to-talk: you hold or press a key to start talking, and you press Enter to send. That is two physical acts on every single prompt. A hundred prompts in a day is two hundred keypresses that exist only to start and stop the tool. Infina removes both of them. Nothing is held, nothing is pressed, and the same loop runs on three spoken words.
There are three triggers and there is no fourth. "type" types, "send it" sends, "open" opens and switches apps. The send trigger answers to either wording, "send it" or a plain "enter", whichever you find yourself saying. That is the entire vocabulary: no command language, no grammar, no custom scripts to write, nothing to memorize beyond those three words.
When Infina is not the answer. If you write in a language other than English, the base product is English only, and Wispr Flow ships 100+ languages.
Everywhere else, the difference is the loop. Every other tool in that table types words into a box and stops, leaving your hands to press Enter and pick the next window. Infina finishes it: say "type" and your sentence, say "send it" (or just "enter"), say "open Cursor", and keep going without a key.