Point. Listen. — Read any on-screen text aloud on macOS
Fully offline · Nothing ever leaves your Mac · Auto voice switching for mixed languages
Download ·
How to use ·
Use cases ·
Privacy
Voice Reader is a point-to-read tool for macOS: any text on your screen — web pages, chats, PDFs, images, video subtitles, scanned documents — can be selected, captured, or simply hovered over to hear it read aloud in natural voices.
It is not a full screen reader like VoiceOver. It is a lightweight "read exactly what I point at" player: a small draggable pill floats on top of everything, and expands into a full player with karaoke-style highlighting when you want to follow along.
| Feature | What it does | |
|---|---|---|
| 🖱️ | Select to speak | Select text in any app — a floating 🔊 button appears next to it; click to listen. Optional "read immediately on selection" mode |
| ⛶ | Region capture (OCR) | Screenshot-style interaction: screen dims → drag a box → text is recognized → preview the image and text → press play. Works on images, scanned PDFs, subtitles |
| ♿ | Hover to read | Toggle it on and the cursor becomes a reading finger: rest it briefly and it speaks the paragraph / sentence / button you point at, starting from the sentence under the cursor. Move away and it stops; each new block gets a soft chime |
| 🗣️ | Smart voices | Language is fully automatic: mixed Chinese / English / Japanese / Korean / Russian text switches voices segment by segment. One-tap female/male voice; per-language voice picks in Settings |
| 📖 | Follow-along highlight | Expand the window to read while listening: the current word is highlighted and auto-scrolled; skip by sentence; drag the progress bar (snaps to sentence start) |
| 🎚️ | Playback comfort | 0.8–2.0× speed (changes apply instantly mid-read) · double-press Stop to clear · all hotkeys customizable |
| 🌐 | Multilingual | UI in English / 简体中文 / 日本語 / Español / Français / 한국어; OCR recognizes 16 languages (CJK, Cyrillic, Thai, Vietnamese, major European languages) |
| 🧭 | Guided onboarding | A 6-step first-run tour walks through permissions with one-click grant buttons and resumes automatically after the app restarts |
- Download the zip from Releases, unzip, double-click
Voice Reader.app - Follow the onboarding to grant two permissions (each row has an "Open" button; the app restarts automatically once granted):
- Accessibility — to read the text you select and respond to global hotkeys
- Screen Recording — to recognize text inside regions you capture
- Done — a pill-shaped mini bar appears in the corner of your screen. That's home base.
- Select: highlight text in any app → click the floating 🔊 (or press
⌃⌥R) - Capture: click ⛶ on the mini bar (or press
⌃⌥S) → screen dims → drag a box around the text → confirm in the preview → ▶ play;Esc/ right-click cancels - Type: expand the window, type or paste with
⌘V, place the cursor where you want to start, press ▶
Turn on the ♿ toggle on the mini bar:
- Rest the cursor for half a second → it reads what you point at: body text is read by paragraph (starting from the sentence under your cursor); buttons and links are announced together with their role ("Save, button")
- Move away to stop; come back and it reads again. Jittering inside the same paragraph never re-triggers
- A soft chime marks each new block
Mini bar (always on top, draggable anywhere, floats over full-screen apps):
▶ play/pause · ⏹ stop · ⛶ capture · ♿ hover-read · 1.0× speed · ⤢ expand
Full window (press ⤢): live highlight over the text, progress bar, ⏮ previous / ⏭ next sentence, quick voice style switch (Auto / Female / Male). Closing it collapses back to the mini bar without interrupting playback.
| Default | Action |
|---|---|
⌃⌥R |
Read the current selection |
⌃⌥S |
Capture a region and read it |
⌃⌥X |
Stop (press twice quickly to stop and clear) |
For voices closer to human: System Settings → Accessibility → Spoken Content → Manage Voices, download the "Enhanced / Premium" voices — Voice Reader picks the best quality automatically.
- Long reads, tired eyes — switch a deep-dive article or paper to 1.5× and listen while you do something else
- Text you can't select — announcements inside images, scanned PDFs, lecture-video slides, long chat screenshots: draw a box, hear it
- Mixed-language material — technical docs and bilingual articles switch between native-sounding voices per segment, no more robotic accents
- Moments when reading is hard — eye strain, small fonts, bad lighting: turn on hover-read and point at whatever you want spoken
- Language learning — replay a sentence at 0.8×, follow the highlighted words, mimic the pronunciation
- Proofreading your own writing — hearing your email or essay read back exposes clumsy sentences instantly
- Hands busy, eyes elsewhere — cooking, stretching, tidying up: turn what you meant to read into something you can hear
- Recognition (Apple Vision) and synthesis (AVSpeechSynthesizer) run entirely on your Mac — no network requests, no telemetry, no account
- The two system permissions are used only for the features above; after grabbing a selection, your original clipboard is restored automatically
- In-app feedback goes through your own mail client — you see and edit everything that gets sent
🚧 The first public build is being prepared (Developer ID signing + notarization), and will land in Releases. Requirements: macOS 12+, Apple Silicon.
🧩 A Chrome companion extension (select-to-speak in the browser, works on Windows/Linux Chrome too) is headed to the Chrome Web Store.
In-app: ⚙ Settings → 💬 Feedback (attaches app version and macOS version automatically), or open an issue in this repository.