Your Mac reads to you.

Select text in any app — browser, PDF, mail, terminal — press ⌥⌘L, and a natural voice reads it aloud. Free, open source, 100% local. No cloud, no account, no subscription.

curl -fsSL https://raw.githubusercontent.com/EloiGils/outloud/main/install.sh | bash

macOS + Homebrew · or clone it from GitHub and run ./install.sh

outloud reading selected text aloud
$139/yr
Speechify Premium
2005
what Apple's built-in voices sound like
$0 forever
outloud — open source, runs on your Mac

Everything the paid apps do. Locally.

Built because reading long text on a screen is hard for a lot of us — and the good tools were locked behind subscriptions.

⚡️

Speaks in ~1.7 seconds

A resident daemon keeps the AI voice model warm and synthesizes sentence-by-sentence. Long article? Same instant start.

🌍

Works in every app

One global hotkey. Browser, PDFs, Slack, terminals — including apps that block synthetic keystrokes.

🗣

Natural voices, 9 languages

50+ voices from Kokoro-82M: English, Spanish, French, Italian, Portuguese, Hindi, Japanese, Chinese.

🎛

Live speed & exact pause

Change speed mid-sentence with pitch correction. Pause freezes on the exact word — not "mute and drift".

📜

Reading history

Every reading saved with cached audio. Re-listen instantly from the menu bar or the Spotlight-style search (⌥⌘H).

🔒

Private by design

Your text never leaves your Mac. No accounts, no telemetry, no network calls. MIT licensed — read the code.

Three shortcuts. That's the whole manual.

Plus a 🔊 menu bar icon for history, voices and speed — and a floating HUD with controls while it reads.

⌥ ⌘ Lread the selected text
⌥ ⌘ Kstop
⌥ ⌘ Hsearch your reading history

FAQ

Is it really free? What's the catch?

Really free, MIT licensed, no catch. It runs an open-source AI voice model (Kokoro-82M) on your own Mac, so there are no server costs to pass on to you.

Does my text get sent anywhere?

No. Synthesis happens on your machine. The only network access is downloading the voice model once at install time.

What does it need?

macOS with Homebrew. The installer sets up the rest: a private Python environment with the model (~2 GB), the mpv audio engine, and the Hammerspoon UI layer. Apple Silicon recommended.

How is it so fast if it's a local AI model?

Three tricks: the model stays warm in a resident daemon, text is synthesized sentence-by-sentence streaming into the player, and the first chunk is cut extra short so audio starts while the rest is still generating. Details in the docs.

Can I use it from the terminal / scripts?

Yes: outloud "text", cat file.md | outloud, outloud --history, outloud --last. Full CLI reference in the docs.

How do I uninstall it?

Run ./uninstall.sh from the repo. It removes everything except your reading history folder, which it tells you how to delete too.