Text to speech for language learners on a Mac

Short answerYour Mac can read text aloud in many languages for free. Open System Settings → Accessibility → Spoken Content (called Read & Speak on newer macOS), use Manage Voices in the System voice menu to download a voice for the language you're learning, select it, and lower the Speaking rate. Then select text and press Option-Esc. In Terminal, say -v '?' lists voices by language. Aloud adds a follow-along reader with word highlight, speed changes mid-sentence (0.75x to 2x) and .m4a export, using the same Mac voices.

Set up a voice for the language you're learning

1. Open the speech settings. Choose Apple menu → System Settings → Accessibility → Spoken Content. On newer macOS the pane may be called Read & Speak.
2. Download a voice for your target language. Click the System voice menu, choose Manage Voices, find the language you are learning and download a voice. Enhanced or Premium voices, where offered, sound more natural.
3. Select it as the System voice. Pick the new voice in the System voice menu and click Play Sample to check it sounds right.
4. Slow the Speaking rate. Drag Speaking rate a little to the left. Slower than normal is easier while you are learning; speed it back up as you improve.
5. Turn on Speak Selection and test it. Switch on Speak selection, select a sentence in your target language in any app and press Option-Esc. Press it again to stop.

Apple has moved and renamed this pane between releases, so if the labels differ on your Mac, search System Settings for "system voice" or "speaking rate". Which languages and voice qualities are offered depends on your macOS version; downloaded voices work offline. If text in your target language is read with an English accent, the English voice is still selected: switch the System voice to your target-language voice. More on setup in how to turn on Speak Selection and how to change the voice and speed.

Tip: switch Highlight content on (in the Speak selection options) so you can see each word as it is spoken, and set Show controller to Automatically to get pause, skip and speed buttons while it reads.

Terminal: hear any sentence in any installed voice

The say command is handy because you can pick a voice per sentence without changing your System voice:

say -v '?'                         # list voices with their language codes
say -v '?' | grep es_              # just the Spanish voices
say -v Monica "¿Dónde está la estación?"
say -v Thomas -r 120 "Je voudrais un café, s'il vous plaît."
say -v Anna -r 120 -f vokabeln.txt

Voice names here are examples; the ones on your Mac may differ, so copy a name exactly as say -v '?' prints it (put names with spaces or brackets in quotes). -r is words per minute; around 175 to 200 is a normal pace, and 110 to 140 is a comfortable slow pace for learning. Add -o phrase.aiff to save the audio to a file (details in export text to audio on Mac).

Practice routines that work

Slow first, then normal speed

• Listen to a short passage at a slow rate while reading along, then again at normal speed.
• Raise the rate a step every few days. Being able to follow at normal speed is the goal, not the slow version.

Shadowing

• Play one sentence, pause, and repeat it out loud copying the rhythm and intonation. Then play it again and compare.
• Keep sentences short. Speak Selection reads whatever you select, so select one sentence at a time.
• Record yourself (QuickTime Player or Voice Memos) now and then to hear your progress.

Listen to real articles in the target language

• Pick news or blog posts slightly above your level. Open Reader view in Safari (or reading mode in Chrome) so only the article is selected, then use Speak Selection. See read web articles aloud on Mac.
• Read along once, then listen again without the text.
• Keep a vocabulary list in a text file and have it read back as a quick review, for example with say -f.

Honest limits

• Synthetic voices make mistakes. Stress, names, numbers, abbreviations and words with several readings can come out wrong, more often in some languages than others. Treat text to speech as a helpful model, not a pronunciation authority, and check important words against native recordings or a teacher.
• Voice quality varies by language. Some languages have natural Premium or Enhanced voices; others may only have a basic, more robotic one.
• Mixed-language text is awkward. A single voice reads everything, so English notes inside a Spanish text (or the reverse) will sound odd. Select only the target-language text.
• No translation or grammar help. Neither the Mac's built-in speech nor Aloud translates; pair them with a dictionary or translator.

What Aloud adds

The built-in tools are fine for a sentence at a time. Aloud (ours, $19 once per Mac) makes longer listening sessions easier:

• A reading window with a word-by-word highlight and click-to-seek, so you can jump back to a sentence you missed and read along.
• Speed from 0.75x to 2x, changed any time, mid-sentence: slow down for a hard paragraph, then speed back up.
• Your Mac's voices, including the target-language Enhanced and Premium voices you downloaded above. (The bundled Kokoro neural voice is English only, with eight voices, on Apple silicon with macOS 15 or later.)
• Read the article on the front tab of Safari, Chrome, Edge or Brave with Control-Option-A, and PDFs without headers, footers and page numbers.
• Read text in images and screenshots (Control-Option-O) with on-device text recognition; how well this works for a given language and script may vary.
• Resume from the last sentence and a sleep timer for longer listening.
• .m4a export to listen to a passage again later, and an optional ElevenLabs or Cartesia voice with your own API key (billed by that provider).

With Apple voices, nothing you read leaves your Mac. Aloud runs on macOS 14 or later. Try your target language in the 3-day trial first; it needs no card or account.