Free alternative to Wispr Flow
Home › Guides · Last checked 2026-10-26 · Facts verified from each project's docs unless marked tested
The closest free alternative to Wispr Flow is Handy, a free MIT-licensed speech-to-text app for Windows, macOS and Linux. Press a keyboard shortcut, speak, release, and the text is pasted into the app you're typing in. Transcription runs entirely on your computer with Whisper or Parakeet V3 models, so there's no weekly word cap and no audio sent to the cloud. The catch: Handy has no phone app, and it transcribes what you say rather than rewriting it into polished prose the way Wispr Flow markets itself.
| Wispr Flow | Handy | |
|---|---|---|
| Cost | Free; Pro $15/user/mo monthly or $12 annually (Wispr Flow pricing page and FAQ) | Free |
| Free limits | 2,000 words a week on desktop; limited weekly use on mobile | None set by the app |
| Offline | Not stated on its pricing page | Yes, fully local |
| Platforms | Mac, Windows, iOS and Android | Windows, macOS and Linux |
| Output | Markets speech turned into polished writing | Your words as transcribed |
| Extras | Dictionary, snippets, meeting Notetaker | Custom shortcuts, model choice, community Raycast extension |
| License | Commercial service | MIT |
Start dictating in 5 steps
- Download the installer for your system from the Handy releases or handy.computer. On Debian or Ubuntu, install the .deb with
sudo apt install ./Handy_*.debso its dependencies come along. - Launch Handy and grant microphone and accessibility permissions.
- Pick a model. Parakeet V3 is the CPU-friendly choice; a Whisper model uses your GPU if you have one.
- Set your shortcut in Settings. By default you hold it to record and release to stop, or switch to tap-to-toggle.
- Click into any text field, hold the shortcut, speak and let go. The text is pasted where your cursor is.
Tips
- If Handy pastes something you copied earlier instead of your words, the README points to Reliable Paste (Beta) or a longer paste delay in the Debug menu (Ctrl+Shift+D, or Cmd+Shift+D on a Mac).
- Behind a proxy or firewall, you can download model files by hand and drop them in Handy's models folder.
- On a Mac with a Bluetooth headset, pick the built-in mic in Handy to avoid the headset dropping to low-quality audio while you record.
The catches
- No mobile: it's desktop only, while Wispr Flow also covers iPhone and Android.
- Raw transcription: Handy's stated goal is one tool, one job, putting what you say into a text box, so expect to clean up filler and punctuation yourself.
- Linux and Wayland: support is limited and needs extra tools for typing and shortcuts.
- Young project: the README says it's actively developed with known issues.
When Wispr Flow is still worth paying for
If you dictate on your phone, want your speech tidied into finished sentences or need team features like a shared dictionary, SSO or HIPAA terms, Wispr Flow does that. Handy wins when you dictate on a computer, go past 2,000 words a week or want dictation that stays on your machine.
Sources: Handy README and GitHub license (MIT), Handy releases (v0.9.8, October 3, 2026), handy.computer, Wispr Flow pricing page and FAQ (all checked 2026-10-26). We have no affiliate link with either. See all free voice tools or read Whisper vs Otter.
FAQ
How much does Wispr Flow cost?
Wispr Flow's pricing page (checked 2026-10-26) lists a Free plan with a 2,000-words-a-week limit on desktop and limited weekly use on mobile. Pro removes the limits for $15 per user a month billed monthly or $12 billed annually. Growth is $23 monthly or $18 annually, and Enterprise is custom. Students and educators get 50% off Pro, per the FAQ.
Is Handy really free?
Yes. The README says Handy is free and open source under the MIT License, and handy.computer says accessibility tooling belongs in everyone's hands, not behind a paywall. The project is funded by sponsors and donations. The Handy name and logo are not open source, so forks must use their own branding.
Does it work offline?
Yes. The README says the process is entirely local: Silero filters silence and your chosen model transcribes on your computer, without sending information to the cloud. You only need the internet to download the app and a model.
Which models can I use?
Whisper Small, Medium, Turbo and Large, with GPU acceleration when available, or Parakeet V3, which the README calls CPU-optimized with automatic language detection. The listed downloads range from 487 MB (Whisper Small) to 1,600 MB (Whisper Turbo); Parakeet is 731 MB. You can also add custom Whisper GGML models.
Which platforms does it run on?
Windows, macOS and Linux. The v0.9.8 release (October 3, 2026) ships Windows installers (x64 and ARM), macOS builds for Apple Silicon and Intel and Linux .deb, .rpm and AppImage packages. Homebrew and winget packages exist but aren't maintained by the Handy developers. There's no iPhone or Android app.
Any setup gotchas?
On macOS you grant microphone and accessibility permissions. On Linux you may need xdotool (X11) or wtype or dotool (Wayland) for typing to work, and on Wayland global shortcuts are set in your desktop environment using the handy --toggle-transcription command. fn/Globe-key shortcuts on a Mac only work on Apple keyboards.