VoiceWink vs Superwhisper
Short answer: For dictation on Windows, choose VoiceWink: a lifetime license from $25, paid once, with every feature in it; the optional cloud features use your own API key. Superwhisper keeps Parakeet, its cloud models, its AI models and your own API keys for Pro, at $8.49 a month, $84.99 a year or $249.99 once. Choose Superwhisper if you also dictate on a Mac or a phone, or want its free tier.
This page compares two dictation apps that run on Windows:
- Superwhisper: a dictation app on Mac, Windows, iOS and Android. Pro costs $8.49 a month, $84.99 a year or $249.99 once; there is also a free tier. Its models can run on your device.
- VoiceWink: a native Windows app with a lifetime license, from $25 paid once, and no subscription. It turns your speech into text on your PC, with your pick of many speech models, and its built-in AI cleans up and rewrites the text for the app you are in. Try it free for 7 days.
Side by side
Green marks the stronger option in a row. Rows without green are about even.
| Feature | Superwhisper | VoiceWink |
|---|---|---|
| Price | A “free tier that doesn’t expire”. Pro: $8.49 a month, $84.99 a year or $249.99 once. | Stronger option. No subscription: a lifetime license from $25, paid once, VAT included. 7-day free trial and 30-day money-back guarantee. Cloud features use your own API key, at the provider’s prices. It is open source, so you can also build the full app yourself, for free. |
| What the price includes | The free tier has Whisper models, custom prompts and meeting recording. Pro adds NVIDIA Parakeet and the cloud models, hosted AI models with no API key needed, your own API keys, translation to English, and audio and video files. | Stronger option. Every feature, in every license: all speech models, AI enhancement on your PC or with your own API key, translation, audio files and voice-to-image. |
| Speech models | Whisper models in the free tier. Pro adds NVIDIA Parakeet, Cohere Transcribe and cloud models from Deepgram, ElevenLabs and Superwhisper. | Your pick of many speech models, made by OpenAI, NVIDIA, Deepgram and ElevenLabs. Some run on your PC; the cloud ones use your own API key. |
| Languages | “100+ languages”, detected automatically. | About 100, depending on the speech model you choose. |
| Works offline | Yes. Superwhisper says: “Your audio never leaves your machine and you don’t need internet.” | Yes, once the models are downloaded, AI enhancement included. A paid license checks in online at least once every 30 days. Cloud features need the internet. |
| Your voice and privacy | Its on-device models keep your audio on your machine. It is “SOC 2 Type II and HIPAA compliant”. | With models on your PC, your speech and text stay there, and so do your recordings and history. With a cloud service, the audio or text goes to the provider or server you chose. No account needed. |
| Cleaning up the text | It “handles punctuation and formatting automatically”. | Adds punctuation and removes filler sounds such as “uh” and “hmm”, in every language. With AI enhancement, it also fixes grammar and removes stutters and false starts. |
| Rewriting for each app | Super Mode reads your screen and adjusts: “Email gets email formatting. Code stays technical. Slack stays casual.” Custom prompts too. | AI enhancement rewrites what you say for the app you are in, for example as an email or a chat message, or translates it. Set it per app, by hotkey or by trigger word, or write your own prompts. It runs on your PC with built-in AI models, or on your own AI server or with a cloud key. |
| AI enhancement on your PC | In Pro, one model, S1-mini: English only, and only in its Message and Mail modes. | Stronger option. Built-in AI models run every text prompt on your PC, your own included. |
| Text while you speak | Stronger option. “Realtime shows results as you record”, with its Nova, Scribe and Parakeet models. | Shows the text after you stop talking. |
| Platforms | Stronger option. Mac, Windows 10 and 11, iOS and Android. | A native Windows 11 app, Arm PCs included. |
| Devices per license | Stronger option. Pro works on “as many of your devices as you like”. | Up to three PCs, depending on the license. |
| Dictionary | Names and terms you enter once: “Superwhisper remembers forever”. | Your names and jargon, plus rules that fix words the engine keeps getting wrong. |
| Audio files | Pro: “Drop in audio or video files and get transcripts”. | Turns audio recordings into text. |
| Meetings | Stronger option. Meeting recording and transcription, in the free tier. | Not in VoiceWink. |
| Voice-to-image | Not in Superwhisper. | Stronger option. Describe a picture and it is made with your own API key, then copied to your clipboard. |
| Source code | Not published. | Stronger option. Open source, on GitHub (GPL v3). |
When Superwhisper is the better choice
- You also dictate on a Mac, an iPhone or an Android phone.
- You want a free tier that does not expire.
- You want one license on as many devices as you like.
- You want your meetings recorded and transcribed.
- You want to see the words while you speak.
When VoiceWink is the better choice
- You would rather pay once, from $25, for a lifetime license, or build it yourself for free from the open source code, than pay $8.49 a month, $84.99 a year or $249.99 once for Superwhisper Pro.
- You want every speech model, your own API keys, AI enhancement on your PC and audio file transcription in every license, not in a Pro plan.
- You want AI on your PC that runs every text prompt, your own included.
- You want voice-to-image.
- You want to be able to read the source code.
Sources
This comparison is by DV Studio, the maker of VoiceWink. Superwhisper’s facts come from four of its own pages, read on 30 September 2026: superwhisper.com, superwhisper.com/windows, its voice models page and its language models page. Features and prices may have changed since. “Not in …” means the product’s own pages, read on that date, describe no such feature. Superwhisper is a trademark of its owner. VoiceWink is not affiliated with, endorsed by or sponsored by Superwhisper’s maker.