BetterWispr
BetterWispr is a free, open-source dictation app for macOS under Apache-2.0, made by opennookorg. You hold Option Space, speak, and the text lands in the app you were typing in. Speech-to-text runs on your Mac with Whisper, NVIDIA Parakeet or Apple on-device recognition, so your voice stays on the computer unless you pick a cloud model yourself. One caveat: macOS only, with Apple Silicon recommended.
What does BetterWispr do?
BetterWispr describes itself as "a private, offline speech-to-text app and an open-source alternative to Wispr Flow." You hold Option Space, speak, release, and the text appears in whatever app you were typing in. The speech model runs on your Mac.
The README lists Apple's on-device speech, NVIDIA Parakeet v2 for English and v3 for 25 European languages, and Whisper large-v3 turbo. Beyond the basic dictation loop, the app has a vocabulary for spelling hints and phrase replacements, a corrections flow that learns misheard words after you fix them once, spoken commands such as "comma", "new line" and "scratch that", a meeting mode with a live transcript and on-device summaries, and a file-and-link transcription panel that uses yt-dlp (installed on demand). An MCP server lets other apps read the meeting notes you choose, read-only.
Key facts
- Licence: Apache-2.0 on the code. Third-party code and model weights keep their own terms.
- Models: Apple on-device speech, NVIDIA Parakeet v2 (English) and v3 (25 European languages), and Whisper large-v3 turbo.
- Platform: macOS 14 or later, Apple Silicon recommended.
- Bring your own model: optional Sarvam AI, Smallest AI, or any OpenAI-compatible endpoint. HTTPS required except on localhost.
- Keys stored in the macOS Keychain. The app never falls back to a cloud provider on its own.
- Audio is written to temporary files and deleted after each recording; settings and history live in
~/Library/Application Support/BetterWispr/. - Install: a DMG from GitHub Releases, dragged to Applications. Microphone and Accessibility permissions required.
- Recordings stop automatically after ten minutes, with a 30-second countdown.
What does it replace, and where does it fall short?
BetterWispr replaces Wispr Flow for the core job: hold a key, speak, paste where the cursor is, on a Mac, with the speech model running on the machine. The README says "your voice never leaves your computer unless you explicitly select an API connection or Codex subscription transcription."
Where it falls short: it is macOS only, with Apple Silicon recommended, so Windows and Linux users have to look elsewhere. The project is in early releases, so expect rough edges. If paste cannot safely reach the previous app the text goes to the clipboard instead. Bring-your-own cloud models follow their own pricing and retention, which the README states but does not audit. "Send to Claude" is off by default; the README documents what it reads about running processes when you turn it on.
How does BetterWispr run?
Download the DMG from GitHub Releases, drag the app to Applications, and open it. The welcome guide picks a model and asks for Microphone and Accessibility permission. Apple on-device models need no download; Parakeet and Whisper download once and then work offline. The app is distributed as a notarised macOS build.
For a self-hosted or remote model, open Models, Bring your own model, Add connection, pick a provider, paste the key, and click Use. Sarvam AI and Smallest AI are listed as built-in options; any OpenAI-compatible endpoint also works. There is no paid tier from the maintainers.
Who is BetterWispr for?
Mac users who want a dictation key that runs on the machine, with the option to try a hosted model later. Someone on Windows or Linux has to pick another entry on the open-source alternatives to Wispr Flow list; Handy covers all three desktops.
What limits does the README state?
Apple on-device models may be unavailable if the language or OS speech assets do not support it; the README suggests installing Parakeet or Whisper instead. First transcription is slow because the model has to load. If paste stops working after an update or a rebuild, the README says to remove and re-add BetterWispr in System Settings, Privacy and Security, Accessibility. Settings, vocabulary, history and voice profiles are stored unencrypted on disk. No tagged release is linked from the README; the download points at Releases.
Questions people ask
Is BetterWispr open source?
Yes. The LICENSE is Apache-2.0, an OSI approved licence, on the code. The README notes that third-party code and model weights keep their own terms, which is standard for a project that bundles Whisper, Parakeet and Apple speech. There is no "additional conditions" clause on the code itself.
Can BetterWispr be used commercially?
Yes. Apache-2.0 allows commercial use, modification and redistribution, provided the licence and the notice file travel with the code. Model weights are a separate matter: Apple's on-device speech is covered by macOS, Parakeet weights are published by NVIDIA, and Whisper weights by OpenAI. Each has its own terms.
Does BetterWispr work offline?
Yes, with built-in models. Apple on-device needs no download. Parakeet and Whisper download once and then run on the Mac. The README states the built-in models only touch the network when you explicitly download one. When you pick a Bring-your-own-model connection, audio goes to that endpoint and its pricing and retention apply.
How does BetterWispr compare with Wispr Flow?
BetterWispr does the core of Wispr Flow, speak and paste, with no subscription and no account, on macOS, with the speech model on the Mac. Wispr Flow is a hosted product; its polished prose and across-device sync are not what BetterWispr sells. The open-source alternatives to Wispr Flow page sets it beside Handy, VoiceInk, SpeakOfLow and OpenWhispr.
Does BetterWispr work on Windows or Linux?
No. The README states macOS 14 or later, Apple Silicon recommended. For Windows and Linux dictation, see Handy on the Wispr Flow alternatives page; it ships installers for all three desktops.
Sources
- BetterWispr README and LICENSE: github.com/opennookorg/betterwispr, read 2026-10-26. The README is the source for every feature, platform note and privacy statement above.
- Project website: betterwispr.com, read 2026-10-26.
More voice dictation apps
A free, offline push-to-talk dictation app for macOS, Windows and Linux that runs Whisper or NVIDIA Parakeet on your own machine and pastes the text at your cursor, under MIT.
A GPL-3.0 macOS dictation app that runs Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper on the device, with command and write modes, per-app prompts, and an optional closed local enhancement model.
A dictation and meeting-notes app under MIT that runs Whisper or Parakeet locally or cloud models with your own key, detects Zoom, Teams and FaceTime calls, and labels speakers on device.
A macOS dictation app under GPL-3.0 whose README names Superwhisper and Wispr Flow as the tools it replaces, with local models, per-app modes, a personal dictionary and an AI assistant mode.
A speech-to-text app inside the Epicenter monorepo under AGPL-3.0 that records, transcribes with a provider you choose, cloud or local, optionally polishes the text, and pastes it at the cursor.
An Apache-2.0 macOS dictation app that calls Gemini 3.5 Transcribe with your own key, writes audio to disk from the first millisecond for crash safety, and inserts polished text at your cursor without an account.
Added October 26, 2026. Every claim above comes from the project's README, LICENSE or model card, read on October 26, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/opennookorg/betterwispr.
Want BetterWispr running on your own servers?
Reveneau sets it up on your hardware, connects it to your product, and builds the software around it.