Talk. It types itself.
Voice to text in any Mac app. Hold ⌥, speak, let go.
Download for macOSfree · macOS 14+ · ~4 MB
One gesture. That's all.
Hold
the right Option key
Speak
any language, auto-detected
Done
text lands at your cursor
Watch a reply appear.
Small, fast, out of your way.
Instant
~0.2s to transcribe — Groq Whisper on an edge network.
Any app
If a cursor blinks there, Lispr types there.
Any language
~99 languages, detected automatically. Switch mid-sentence.
Menu bar
~4 MB, no window, nothing to configure.
Clipboard-safe
Inserts your text, then restores what you'd copied.
Push to talk
Hold, speak, release. No wake word, no listening mode.
~99 languages — Lispr detects which on its own.
You think faster than you type.
Where Lispr fits.
| Lispr | Wispr Flow | Apple Dictation | |
|---|---|---|---|
| Price | Free | $15/mo | Free |
| No subscription | ✓ | — | ✓ |
| No account or sign-up | ✓ | — | ✓ |
| Languages, auto-detected | ~99 | ~100 | manual |
| Push-to-talk, one key | ✓ | ✓ | — |
| Footprint | ~4 MB, menu bar | full app | built-in |
| Speech model | Whisper large-v3 | proprietary | on-device |
Public details as of May 2026. Lispr is in early access — it does one job, fast.
Said, transcribed, forgotten.
Your audio is sent over an encrypted edge proxy purely to be transcribed — then discarded. Nothing is stored on a server. Nothing trains a model.
The honest answers.
What permissions does Lispr need?
Three, granted once in System Settings: Microphone (to hear you), Accessibility (to paste text into other apps), and Input Monitoring (to catch the ⌥ key).
Which Macs are supported?
macOS 14 Sonoma or newer, on Apple Silicon and Intel. The whole app is around 4 MB.
Does it work offline?
No — transcription runs in the cloud, which is why it's this fast and accurate. The round trip is about 200 ms.
Which languages does it understand?
Around 99. Lispr transcribes with the Whisper speech model, which is multilingual — the language is detected automatically, and you can even switch within a sentence.
Is my voice stored anywhere?
No. Audio is transcribed and immediately discarded. Nothing is kept on a server, nothing trains a model.
Is it really free?
Yes — Lispr is free while it's in early access.