Text to Speech on Mac OS X: What Still Works Today

Short answer: text to speech on Mac OS X still works, including on old Intel Macs, because the speech engine installs with the operating system and is not a separate download. On older releases you read on-screen text with the accessibility pane (called Spoken Content on Big Sur 11.0 and earlier, and Read and Speak on current macOS), or you drive the built-in say command from Terminal. Both still exist. What the built-ins cannot do is clone a voice, design a new one, or run a multi-speaker project from a saved, editable script.

This guide is for anyone on an older Mac, or an old macOS version, who wants to know which text to speech features still function and at what point it is worth moving to a local AI voice app. It sticks to what Apple documents today, names the tools you actually have, and is honest about where an old machine stops being the right place to do the work.

What still works on an older Mac OS X system

Almost everything basic still runs. Apple still publishes the instructions for the on-screen speaking feature all the way back to macOS Mojave 10.14, and the engine that turns text into audio has been part of the system far longer than that. You do not need an account, a subscription, or an app install to make a Mac talk. The pieces below are the ones most people forget are already there.

• The say command in Terminal, which Apple's own man page describes as a tool that converts text to audible speech and either plays it or saves it to a file.
• Speak selection, the keyboard shortcut that reads the selected text or the available text in the current window. The default is Option plus Escape.
• The system voice list, including voices you download from Apple later.
• Pronunciations, so a name, brand, or acronym is spoken the way you want instead of being mangled by a general purpose voice.
• AppleScript's own say command, which is the scripting route into the same speech engine.

Use the say command for speech and for audio files

The say tool is the oldest and most flexible built-in option. Apple's man page describes it plainly: it converts text to audible speech and either plays it through the sound output device chosen in System Preferences, or writes it to a file. The version Apple published for Mac OS X 10.6 already supported the core flags shown here, and the same command still works on current macOS. Open Terminal and try a few of these.

• Say a phrase out loud: type say Hello, World and press Return.
• Speak a file instead of typing: say -f notes.txt reads the file aloud.
• Pick a voice: say -v Alex says the text with the Alex voice.
• Save to audio: say -o hello.aiff Hello, World writes an AIFF file, which is the default output format.
• Request another format on a newer release: for example say -o hi.caf --file-format=caff Hi. Apple's man page notes that formats other than AIFF were introduced in Mac OS X 10.6, so on an older system plan around AIFF.

If you script on a Mac, you do not even need the shell. Apple's Mac Automation Scripting Guide shows the AppleScript form, where say can name a voice and set rate, pitch, and modulation in one line, for example: say "Processing is complete." using "Alex" speaking rate 140 pitch 42 modulation 60. The same guide shows the saving as parameter, which writes the spoken text to an aiff file. That is the practical way to turn an email or a note into audio you can play or archive later without leaving the Mac.

The catch is that say is a narrator of one voice reading a flat script. You can change rate, pitch, and modulation, and you can loop it in a shell script across a folder of text files, but you cannot hand different paragraphs to different speakers, you cannot clone a voice from a sample, and the result will not sound like a produced audiobook. For a paragraph or a utility message that is completely fine. For a client deliverable it usually is not.

Spoken Content became Read and Speak: find the pane on your version

The on-screen reading feature has kept the same behavior while the settings around it were renamed and reorganised. On macOS Big Sur 11.0 and earlier, Apple documents the controls under System Preferences, then Accessibility, then Spoken Content, with a Speak selection checkbox. On current macOS the settings app is System Settings and the pane is Read and Speak, with Speak selection turned on in the same place. The default shortcut has stayed at Option plus Escape across those releases, which is why advice written years ago still applies to a current machine.

1. Open the Apple menu.
2. On Big Sur 11.0 or earlier, open System Preferences, click Accessibility, then click Spoken Content. On current macOS, open System Settings, click Accessibility, then click Read and Speak.
3. Turn on Speak selection.
4. Select text in any app and press Option plus Escape to hear it. If nothing is selected, your Mac speaks the available text items in the current window.
5. Open the Options or info button to change the shortcut, turn on word and sentence highlighting, or show the onscreen controller that can change speaking rate and skip by sentence.
6. Use the System voice pop-up, then its info or Manage Voices button, to add or sample a voice. Downloads need an internet connection, and Apple notes the voice is not available until the download finishes.

Two more settings do a lot of work once a voice is in place. The system speech language sets which language your Mac uses to speak, and Pronunciations lets you define how a specific word should be said, either by dictating it or spelling it out. If your Mac will not speak selected text, the usual causes are that nothing readable was selected, that the app does not hand its text to the speech system, or that another app has captured the shortcut. Text baked into a scanned image or an image-only PDF will not read until it has been recognized first.

Choose the right route for the job

The table below is a fit guide, not a quality ranking. The built-in routes win on cost and speed and are right for most listening. The lower rows only become relevant when the output has to become a file, has to hold a consistent or custom voice, or has to survive a review and edit loop.

For a wider view of the reader apps and studios that run on a Mac, see the guide to the best text to speech apps for Mac. It separates reading tools from production tools so you can tell which category you actually need before you spend anything.

Where an older Mac runs out of road

An old Mac is not only an older operating system. Applications built for a previous macOS can stop launching once their 32-bit support is gone, and the newest AI voice models are built for current graphics and unified memory. This is the honest boundary: the built-in text to speech keeps working, but the newest generative voices on your machine often need hardware and an operating system that an old Mac does not have.

Murmur is a local voice studio for Apple Silicon Macs running macOS 15 or later, so it will not install on an old Intel Mac or an old macOS release. If that is your machine, your text to speech answer is the built-in one plus, if you want a library and long-form reading, a reader app. Murmur is simply not an option for you, and it would be misleading to suggest otherwise. If you are on a recent Apple Silicon Mac, Murmur runs its models on the machine, so text and audio stay local, there are no per character fees, and nothing is uploaded. It handles multiple speakers, voice cloning from a reference sample, and voice design, and it exports audio from a queue.

The pricing is a one time $49 with a free plan and a 3-day Pro trial, and a 7 day refund policy, so the sensible move is to listen to the public samples before buying anything. If you only need the built-in tools, keep them. The path from a reading session to a finished file is described in the Mac text to speech workflow, and the shortcut and automation detail is in the Mac text to speech shortcut guide. Readers weighing a one time app against a subscription can see the trade offs in one time TTS app versus subscription.

A short checklist before you commit

• Just need to hear text now? Use Speak selection and stop there.
• Need a quick audio file from a short string? Use say -o in Terminal, and expect AIFF unless your release supports more formats.
• Need a nicer voice on a recent Mac? Add or download a system voice first. It is free.
• Need a specific or personal voice? A system voice will not do it. That requires cloning or voice design.
• Need several speakers in one project? Use a tool with multi-speaker editing, because say reads one voice at a time.
• On Windows, a phone, or a browser, or you need live streaming? Neither the built-ins nor Murmur cover that, so choose a platform that does.
• Working with private documents? Keep it on the machine, and see the offline text to speech for private Mac documents guide for the workflow.

Frequently asked questions