The open-source alternatives to Wispr Flow are Handy, VoiceInk, OpenWhispr and Whispering. All four do the same core job: hold a key, speak, and the text lands at your cursor. Handy and OpenWhispr run on macOS, Windows and Linux; VoiceInk is macOS only. Each can run the speech model on your own machine, using Whisper or NVIDIA's Parakeet, so the audio never leaves it. OpenWhispr and VoiceInk add an optional language-model step to clean up the text. A team switches when it wants dictation with no subscription, no cloud, and a licence it can read.
15 of 15 carry an OSI licence. 15 of 15 run on your own machines. Figures read from GitHub on October 26, 2026.
One project per need, with the reason from its own README and the one thing to know before you choose it.
MIT, offline, and one job: hold a key, speak, paste. Installers for macOS, Windows and Linux, with Whisper on a GPU or NVIDIA's Parakeet on a CPU.
Know firstIts post-processing step is a command-line toggle whose behaviour the README does not describe, so polished prose is not the point here.
GPL-3.0, offline, with modes that change settings per app or site, a personal dictionary, and an AI assistant mode. Its own README names Superwhisper and Wispr Flow as the tools it replaces.
Know firstmacOS 14.4 or later only, and the project is not accepting pull requests.
GPL-3.0, on-device Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper, with a write mode that rewrites selected text and a command mode.
Know firstmacOS 15 or later and Apple Silicon for all models; the optional enhancement model, Fluid Intelligence, is local but closed.
AGPL-3.0, seven bundled local models including Nemotron 3.5 and Qwen3-ASR, cleanup through Ollama or any OpenAI-compatible endpoint, and a self-hosted server option.
Know firstWindows only for the client, and the community channels are Chinese-language by the README's own note.
MIT, built on Handy's dictation core, with a trigger phrase, a local English cleanup model, screen-aware answers and Kokoro speech, on all three desktops.
Know firstUnsigned installers that Windows and macOS warn about, and a young project at its first releases.
MIT. Local models or cloud models with your key, a cleanup step through GPT, Claude, Gemini, Groq or a local model, and call detection with speaker labels on device.
Know firstNotes sync and team spaces run through the maintainers' hosted cloud.
AGPL-3.0, inside the Epicenter monorepo. Any cloud or self-hosted transcription endpoint, an optional polish step, and a local-first store you own.
Know firstBuilt from source with Bun; neither README lists an installer to download.
The facts that decide most choices. Each project's page carries the full panel.
| Project | Best for | Openness | Licence | Runs as | Own machines | Stars |
|---|---|---|---|---|---|---|
| Handy | All three desktops, nothing else running | OSI | MIT | Desktop app | Yes | 34,761 |
| VoiceInk | A Mac, with modes per application | OSI | See LICENSE | Desktop app | Yes | 6,933 |
| FluidVoice | A Mac with the newest local models | OSI | GPL-3.0 | Desktop app | Yes | 12,462 |
| SayIt | Windows, with an editable cleanup step | OSI | AGPL-3.0 | Desktop app, Docker (self-hosted server) | Yes | 453 |
| SpeakoFlow | Handy plus a cleanup model and an assistant | OSI | MIT | Desktop app | Yes | 366 |
| OpenWhispr | Cleanup by a model you choose, plus meetings | OSI | MIT | Desktop app | Yes | 10,041 |
| Whispering | Choosing your own transcription provider | OSI | See LICENSE | Built from source (Bun), Desktop app (Tauri), Browser | Yes | 4,826 |
| Jot | OSI | Apache-2.0 | Desktop app | Yes | 497 | |
| BetterWispr | OSI | Apache-2.0 | Desktop app | Yes | 459 | |
| Whispree | OSI | MIT | Menu-bar app | Yes | 131 | |
| yap | OSI | MIT | pipx, from source, or PyInstaller-frozen app | Yes | 77 | |
| hyprwhspr-rs | OSI | MIT | Rust binary (daemon), systemd service, Nix package | Yes | 35 | |
| Voice Gecko | OSI | MIT | Desktop app (Tauri) on Windows, with macOS and Linux builds "when available" | Yes | 35 | |
| OpenFlow | OSI | MIT | Desktop app | Yes | 15 | |
| sayszz | OSI | MIT | Python package, menu-bar app | Yes | 12 |
Licence and star figures read from GitHub on October 26, 2026. Stars are shown as a dated fact and were not used to rank this page.
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.
MIT, offline, and one job: hold a key, speak, paste. Installers for macOS, Windows and Linux, with Whisper on a GPU or NVIDIA's Parakeet on a CPU.
Know firstIts post-processing step is a command-line toggle whose behaviour the README does not describe, so polished prose is not the point here.
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.
GPL-3.0, offline, with modes that change settings per app or site, a personal dictionary, and an AI assistant mode. Its own README names Superwhisper and Wispr Flow as the tools it replaces.
Know firstmacOS 14.4 or later only, and the project is not accepting pull requests.
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.
GPL-3.0, on-device Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper, with a write mode that rewrites selected text and a command mode.
Know firstmacOS 15 or later and Apple Silicon for all models; the optional enhancement model, Fluid Intelligence, is local but closed.
An AGPL-3.0 voice typing app for Windows that ships seven local speech models including Nemotron 3.5 ASR and three Qwen3-ASR sizes, with an editable AI cleanup step through Ollama or any OpenAI-compatible endpoint.
AGPL-3.0, seven bundled local models including Nemotron 3.5 and Qwen3-ASR, cleanup through Ollama or any OpenAI-compatible endpoint, and a self-hosted server option.
Know firstWindows only for the client, and the community channels are Chinese-language by the README's own note.
An MIT desktop voice assistant for Windows, macOS and Linux that extends Handy's dictation core with a trigger phrase, a local cleanup model, screen-aware answers and Kokoro speech, with no account and no telemetry.
MIT, built on Handy's dictation core, with a trigger phrase, a local English cleanup model, screen-aware answers and Kokoro speech, on all three desktops.
Know firstUnsigned installers that Windows and macOS warn about, and a young project at its first releases.
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.
MIT. Local models or cloud models with your key, a cleanup step through GPT, Claude, Gemini, Groq or a local model, and call detection with speaker labels on device.
Know firstNotes sync and team spaces run through the maintainers' hosted cloud.
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.
AGPL-3.0, inside the Epicenter monorepo. Any cloud or self-hosted transcription endpoint, an optional polish step, and a local-first store you own.
Know firstBuilt from source with Bun; neither README lists an installer to download.
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.
A free, open-source dictation app for macOS under Apache-2.0 that holds a shortcut, listens, and pastes the text where your cursor is, with Whisper, NVIDIA Parakeet or Apple on-device speech recognition.
A free, MIT dictation app for macOS that speaks into any prompt box on your Mac, uses Groq for speech and a Codex OAuth token for language-model cleanup, and can also run local Whisper and Qwen models.
A free, MIT push-to-talk dictation app for macOS, Windows and Linux that runs faster-whisper locally by default, learns your vocabulary, and can also send audio to Groq or an OpenAI-compatible endpoint.
An MIT-licensed Rust rewrite of hyprwhspr, built as a background daemon that listens for Hyprland shortcuts, transcribes with whisper.cpp or Parakeet locally, or with Groq, Gemini or an OpenAI-compatible endpoint.
A free, MIT dictation desktop app that runs Parakeet, Moonshine or Whisper on your own machine, ships an installer for Windows today, and has an optional paid Support tier that unlocks no features.
An MIT-licensed desktop dictation app forked from Handy, with three independent backend modes for speech-to-text and AI cleanup, a wake-word mode and OS-keychain API keys.
An MIT-licensed on-device dictation app for macOS and Windows that tap-triggers Whisper, with Metal acceleration on Apple Silicon through MLX, a searchable local history and a setup wizard.
The questions that settle it, in the order they usually come up.
macOS only: VoiceInk. Windows or Linux as well: Handy or OpenWhispr. All three ship installers.
OpenWhispr, with the cleanup model you choose, or VoiceInk with its assistant and context modes. Handy pastes what you said.
Handy and VoiceInk are offline only. OpenWhispr is offline when you choose local models. Whispering depends on the provider you configure.
Wispr Flow, in the vendor's own words, is "the voice-to-text AI that turns speech into clear, polished writing in every app", by its own description. wisprflow.ai. The cases below are where that still wins.
Stars, licence, last commit and last release are read from the GitHub API by a script every week and carry the date they were read. Nobody types them.
What each project does is written from its own README and LICENSE, read in full on the date shown, and phrased as the project's claim. The sources are listed at the end.
That a project is faster, better or cheaper than Wispr Flow. A price for anyone. A user count. A roadmap. If a fact is not on this page, it was not in the source.
For macOS, Windows and Linux with everything local, Handy (MIT) or OpenWhispr (MIT). For macOS only with a language-model cleanup step and a personal dictionary, VoiceInk (GPL-3.0). Whispering (AGPL-3.0) lets you pick the transcription provider, cloud or local, and is built from source. All four paste the text at your cursor after a hotkey.
Handy and VoiceInk run the speech model on your computer only, and both READMEs say no audio leaves the device. OpenWhispr offers both: local models (Whisper, Parakeet and others) or cloud models with your own API key. Whispering sends audio only when the transcription provider you chose needs an upload.
Handy ships installers for macOS, x64 Windows and x64 Linux. OpenWhispr ships .dmg, .exe, .AppImage, .deb and .rpm files. VoiceInk is macOS 14.4 or later only. Whispering is built from source with Bun and runs as a desktop app or in a browser.
VoiceInk and OpenWhispr both run the transcript through a language model you choose, which is where the cleanup happens; OpenWhispr names GPT, Claude, Gemini, Groq and local models as options. Handy has a post-processing toggle in its command line, and its README does not say what the step does. Whispering describes an optional polish step.
All four are built on Whisper, and three also support NVIDIA's Parakeet: Handy (Parakeet V2 and V3), VoiceInk (Parakeet through FluidAudio, plus SenseVoice Small) and OpenWhispr (Parakeet plus Cohere Transcribe). Parakeet is the CPU-friendly option in Handy, which lists an Intel 6th generation or equivalent as the minimum.
Yes, VoiceInk is built for macOS and is GPL-3.0; you can build it from source or install it through Homebrew, and the maintainers also sell a compiled build with automatic updates. Handy and OpenWhispr both run on Intel and Apple Silicon Macs too.
Reveneau sets up the replacement on your own servers, connects it to your product, and builds the software around it.