LerroDownload previewDownload
Public preview for macOS

Speak freely.
Write clearly.

Lerro turns your voice into clean, editable text in any Mac app— right where your cursor is waiting.

Download for macOS View source

Free · Open source · macOS 26+ · Apple silicon

Launch notes

Lerro turns a rough thought into clear text—right where the cursor is.

Local-firstBring your own keyNo accountOpen source

One shortcut, three modes

Your voice can write, translate, or think with you.

Dictate

Speak naturally. Lerro cleans up the sentence and places it at your cursor.

Translate

Say it once, then write it in the language your conversation needs.

Ask

Select text, ask a question, and turn the answer into your next edit.

Choose the right intelligence

Fast by default. Smarter when you want it.

Keep a direct Apple Speech path, bring a fast cloud model, or run the optional model locally.

Raw

Apple Speech

Direct transcription with the shortest path from voice to cursor.

BYOK

Your API key

Use DeepSeek or another OpenAI-compatible provider for fast refinement.

Recommended
Local

Qwen on your Mac

Keep model processing offline when local control matters most.

Permission with purpose

Every permission has one clear job.

Lerro explains each macOS permission before asking. Audio saving stays off by default, and model context remains under your control.

Read the privacy model
01

Microphone

Captures only the voice you choose to dictate.

02

Speech Recognition

Turns audio into the raw words Lerro works with.

03

Accessibility

Returns the finished text to the active app.

04

Input Monitoring

Lets your chosen shortcut work across macOS.

Three steps

Ready whenever the cursor is.

  1. 1

    Choose a shortcut

    Use Fn, another modifier, or a combination that feels natural.

  2. 2

    Speak naturally

    The HUD responds immediately, follows your voice, and shows processing at a glance.

  3. 3

    Keep writing

    Your finished text lands at the active cursor, ready to edit or send.

Preview release

Give your keyboard a voice.

Built for macOS 26 or later on Apple silicon.

Open releases GitHub repository
First-launch note

This is a preview build. If macOS blocks it, confirm the download came from the Lerro GitHub repository, then open System Settings → Privacy & Security → Open Anyway.

Questions, answered

Before your first dictation.

Is Lerro free?

Yes. Lerro is an open-source macOS app. You can use raw dictation for free, bring your own model API key, or download the optional local model.

Where does my voice data go?

Raw transcription uses Apple Speech. Local mode keeps model processing on your Mac. BYOK mode sends the transcript and the context options you enable to your selected provider.

Why does Lerro need Accessibility and Input Monitoring?

Accessibility lets Lerro paste finished text at the active cursor. Input Monitoring lets a single key or shortcut start dictation from any app. The onboarding flow explains each permission before you grant it.

Which Macs are supported?

The current preview requires macOS 26 or later on Apple silicon. Local AI also needs roughly 3.03 GB for the optional model download.

Why might macOS block the preview the first time?

The current public build is a preview. If macOS blocks the first launch, open System Settings, choose Privacy & Security, and use Open Anyway after confirming the download came from the Lerro GitHub repository.