TL;DR
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
BetterWispr has been presented in a Show HN launch as a free, open-source dictation app for macOS 14 and later. It records speech through a keyboard shortcut and transcribes it on the Mac with a selected local model; the project also describes optional cleanup, per-app writing styles and meeting notes.
BetterWispr has launched as a free, open-source dictation app for macOS 14 and later, offering Mac users a way to dictate into other apps using speech models that run locally. The project says dictation does not require an account or cloud transcription service, while Apple Silicon is recommended for local processing.
The app’s basic workflow is to place the cursor in a text field, hold Option + Space while speaking, then release the shortcut to transcribe. BetterWispr says the resulting text is inserted into the app in use; if an app blocks automatic insertion, the transcript can be copied from the app. Transcription starts after recording ends rather than appearing word by word during speech.
BetterWispr lists nine built-in models spanning Apple speech, Parakeet and Whisper. Users can also connect to Sarvam AI, Smallest AI or an OpenAI-compatible API. The site says the local models work offline after their files and required assets have been installed. Model language coverage and performance depend on the selected model and Mac.
Other listed features include optional English cleanup, custom vocabulary hints and replacements, per-app writing styles, voice commands and a history of raw and corrected dictation. A meeting notetaker can transcribe the user’s microphone as “Me” and other app audio as “Them”; the site says Apple Intelligence or Ollama can create summaries, decisions and action items locally. These are product descriptions, not independent performance tests.
Local Dictation Without a Subscription
BetterWispr’s pitch combines local processing with a free, open-source license, addressing two common considerations for people who dictate sensitive or work-related text: where their speech is processed and whether access requires a paid plan. The project says speech stays on the Mac and that the app has no account requirement or cloud fallback for its built-in local transcription.
That does not mean every feature operates without outside services. BetterWispr offers connections to third-party APIs, and its own description says those providers receive dictation when a user chooses an API connection. Users also need to install model files, and some features depend on macOS permissions or local model availability. The distinction matters for anyone evaluating privacy: the app describes a local default, but optional integrations have different data handling.
The Apache 2.0 license lets users inspect and build the software, according to the project. That offers a route to examine how the app works, though the license alone does not verify the site’s privacy or performance claims. The practical value will depend on transcription accuracy, speed and compatibility on individual Macs and with the languages users need.
As an affiliate, we earn on qualifying purchases.
How the Mac App Works
BetterWispr is designed to put dictated text into a user’s current app rather than requiring them to write inside a separate editor. Its site describes a hold-to-talk shortcut, with settings for toggle recording or a different keyboard combination. It also says users must grant microphone permission and Accessibility permission for automatic text insertion; Apple’s speech option requires Speech Recognition permission.
The product offers different cleanup levels: “Light” removes English filler words, repetitions and false starts; “Medium” can also edit for clarity using a chosen notes model; and “None” keeps the spoken wording. The site says styles, cleanup and voice commands are limited to English, while recognition language coverage varies by model. Parakeet’s listed coverage includes English, Japanese and 25 European languages, and Whisper is described as offering broader multilingual recognition.
BetterWispr says dictations can be retained locally as raw and corrected text, with history able to be turned off. Its website also describes updates delivered through GitHub Releases using Sparkle, with installation dependent on a matching EdDSA signature. The supplied launch material does not include independent testing, benchmark results or user adoption figures.
“Speech stays on your Mac.”
— BetterWispr, product website
As an affiliate, we earn on qualifying purchases.
Performance and Privacy Details
The supplied information comes from BetterWispr’s own website; it does not include independent testing of transcription accuracy, latency, resource use or privacy practices. The site labels its timing illustration as not a benchmark, so it should not be read as evidence of measured speed or productivity gains.
Actual results may vary with Mac hardware, operating system assets, model choice, language and recording conditions. The source does not provide a complete list of the nine model names, file sizes, hardware requirements beyond recommending Apple Silicon, or comparative accuracy results. It also does not detail how each optional external API provider handles data. Those points would need to be checked against the app’s settings, documentation and the providers’ policies.
As an affiliate, we earn on qualifying purchases.
Testing the App on Mac
Readers interested in the release can review the source code and build instructions through the project’s GitHub links, or install the app from the channels listed by BetterWispr. The site says setup includes granting relevant permissions and downloading a local speech model. Users should check the selected model’s language support and decide whether to enable dictation history, cleanup features or external API connections.
Further reporting would require hands-on testing across supported Macs and languages, as well as review of the code and any third-party integrations. BetterWispr has not supplied a public benchmark or independent audit in the launch material, and no adoption figures or roadmap details are provided there.
As an affiliate, we earn on qualifying purchases.
Key Questions
Is BetterWispr free?
Yes. BetterWispr says it is free and open source under the Apache 2.0 license, with no dictation subscription, account or API key required for its built-in local workflow. Model files use storage on the Mac.
Does dictation work offline?
According to the project, local transcription works offline after a compatible model and its required assets have been installed. Downloading models requires an internet connection, and Apple speech may depend on supported languages and operating-system assets.
Does BetterWispr transcribe speech live?
No. The site says transcription begins after recording stops; it does not describe words appearing as the user speaks.
Which languages does it support?
Language coverage depends on the chosen model. BetterWispr lists Parakeet options for English, Japanese and 25 European languages, and describes Whisper as having broader multilingual recognition. The site’s cleanup, writing styles and voice commands work in English.
Does it send dictation to an online service?
The project says its built-in local models do not use cloud transcription or a cloud fallback. However, users can select third-party API connections; the site says dictation is sent to that provider when such an option is chosen.
Source: hn
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
