Dictate
Apple Speech uses your app-aware dictionary, while Lerro previews every word and safely writes the final text.
Shortcut→Live preview→Apple Speech→Cursor
Native voice typing for macOS 26
Speak.
Your Mac writes.
Press your shortcut, follow the live transcript, then press again to write. Apple Speech keeps the core path immediate; optional AI adds polish, translation, automatic vocabulary learning, and per-app tone.
Free forever · Apple silicon · macOS 26+
The centered HUD expands smoothly with the live transcript, processes the result, writes at the captured cursor, and disappears after success.
One shortcut, one clear path
See every word.
Straight to the cursor.
Press once to speak, watch the centered preview grow, then press again to finish. A successful delivery ends there; a recovery card appears only when writing fails.
Polish
Your selected local or remote AI applies dictionary terms and the tone saved for the current app.
Apple Speech→Dictionary + tone→Selected AI→Cursor
Translate
Speak naturally and let the selected local or remote AI translate the final transcript before delivery.
Apple Speech→Selected AI→Translation→Cursor
Lerro Personalization
It learns your words. It fits every app.
Correct a recently dictated name or term and AI can add the useful mapping to your app-aware dictionary. Personalization keeps each app's icon, tone, AI preview, and enablement in one first-level page.
Private by architecture
Your voice stays close.
Apple Speech handles the core transcript. Optional intelligence runs through the local MLX model or the remote provider you configure. Audio saving stays off by default.
Read the privacy model- 0
- accounts
- 0
- subscriptions
- 0
- telemetry events
Clear network boundaries
Remote polish and translation send the transcript and enabled context to your provider. Automatic learning sends the smallest recent correction span for classification. Local AI keeps both operations on your Mac.
Open by default
Inspect every product decision.
Lerro is Apache-2.0 licensed. The app, privacy policy, architecture, tests, release scripts, and website are all public.
Questions, answered
Before your first dictation.
Microphone and Speech Recognition power Apple dictation. Accessibility enables the global shortcut, captures the target field, writes finished text, and observes a recent correction when automatic dictionary learning is enabled.
Core Apple dictation works after its language resources are ready. The optional local MLX model keeps polish, translation, app tone, and dictionary learning offline after its 3.03 GB download. Remote AI uses your configured provider.
Apple Speech manages dictation resources at the system level. A compatible resource may already be present; macOS downloads one when the selected language needs it.
Lerro has no account system, subscription, product analytics, or telemetry. Audio saving is off by default. Your configured remote provider receives only the transcript, enabled context, or minimal correction span required for the AI action. Local AI keeps those actions on your Mac.
Lerro requires Apple silicon and macOS 26 or later. The optional local model uses about 3.03 GB of storage and is downloaded only after you approve it.
Every public build is signed with Apple Developer ID, notarized by Apple, and distributed through Lerro's Cloudflare release service. In-app updates verify their signed update metadata before installation.
Lerro 1.6
Speak. Preview. Write. Done.
Signed, notarized, and ready for Apple silicon Macs running macOS 26 or later.