How to Turn On Text to Speech on Windows, Mac, iPhone, Android, and Chromebook

Every major operating system shipped with text to speech built in for years. Windows has it. macOS has it. Your phone has it. But fewer than 12% of users have ever toggled it on, according to accessibility usage surveys. Not because they don't want it, but because the setting is buried under menus most people never open.

The feature can read emails, articles, documents, and entire web pages out loud. Turning it on takes under a minute on any device. Getting a voice that doesn't make you regret the decision takes a bit more thought.

Windows 10 and 11

Windows offers two separate TTS features. Narrator is the full screen reader that announces everything on screen. Read Aloud is a lighter tool built into specific Microsoft apps.

Enabling Narrator

Narrator reads all interface elements: buttons, menus, notifications, and body text. To turn it on:

• Press Win + Ctrl + Enter for instant activation
• Or open Settings > Accessibility > Narrator (Windows 11) or Settings > Ease of Access > Narrator (Windows 10) and flip the toggle

Once active, Narrator starts speaking immediately. A few settings worth adjusting right away:

• Voice selection: Under Narrator settings, click "Choose a voice" to switch between installed options. Microsoft David and Microsoft Zira are the defaults. Additional voices can be downloaded from the same menu
• Speed and pitch: Adjust the speech rate slider. Default is usually too slow for comfortable listening. Start around 60-70% and fine-tune from there
• Verbosity: Narrator can announce every UI detail or just the essentials. Under "Verbosity," reducing the level to 3 or 4 cuts out repetitive announcements like "button" and "checkbox" after every element

Using Read Aloud in Edge and Word

For reading specific content rather than narrating the entire interface, Microsoft Edge and Word both include a Read Aloud feature with better-sounding voices than Narrator.

• In Edge: Open any web page, press Ctrl + Shift + U, or click the three-dot menu and select "Read Aloud." A playback bar appears at the top with voice and speed controls
• In Word: Go to Review > Read Aloud. The feature reads your document from the cursor position forward

Read Aloud uses Microsoft's online neural voices when you're connected to the internet, which sound noticeably more natural than Narrator's offline voices.

macOS

macOS splits TTS into two tiers. Spoken Content reads text on demand. VoiceOver is the full screen reader for navigating the entire interface by voice.

Enabling Spoken Content

This is the option most Mac users want. It reads highlighted text or full screens without changing how your Mac works.

• Open System Settings > Accessibility > Spoken Content
• Toggle on Speak Selection to read any highlighted text via a keyboard shortcut
• Toggle on Speak Screen to read everything visible on the current screen

After enabling Speak Selection, highlight any text and press Option + Esc to hear it read aloud. A small controller appears with play, pause, and speed controls.

Worth configuring:

• System Voice: Click the dropdown to browse available voices. The options labeled "Siri Voice" sound significantly more natural than legacy voices like Alex or Samantha
• Speaking Rate: The default is conservative. Slide it up until the voice sounds conversational, not sluggish
• Show Controller: Enable this to get a persistent playback overlay whenever speech is active

Enabling VoiceOver

VoiceOver is macOS's full screen reader. It announces every interface element and changes how navigation works. Most users don't need VoiceOver unless they rely on it for accessibility.

• Press Cmd + F5 to toggle VoiceOver on
• Or go to System Settings > Accessibility > VoiceOver and flip the toggle

VoiceOver has a learning curve. When it's active, you navigate with keyboard shortcuts rather than mouse clicks, and the system announces each focused element. Apple includes a built-in tutorial accessible from the VoiceOver settings panel.

iPhone and iPad

iOS offers several TTS options, ranging from reading a single highlighted sentence to narrating your entire screen.

Enabling Speak Selection and Speak Screen

These two features cover most use cases without changing how your device works.

• Go to Settings > Accessibility > Spoken Content
• Toggle on Speak Selection: Adds a "Speak" button to the text selection menu. Highlight any text, tap "Speak," and the device reads it aloud
• Toggle on Speak Screen: Swipe down with two fingers from the top of the screen to have the entire page read aloud. A playback controller appears with speed, skip, and pause controls

Additional options in the same menu:

• Highlight Content: Turns this on to see words or sentences highlighted in real time as they're spoken
• Voices: Tap to download enhanced or premium voice packs for your language. Premium voices are larger downloads but sound notably better
• Speaking Rate: Adjustable via slider. Test it with a paragraph of real content rather than the preview sentence

Enabling VoiceOver

VoiceOver on iOS is a full screen reader that changes touch gestures. Single-tap selects an item and reads it aloud. Double-tap activates it.

• Go to Settings > Accessibility > VoiceOver and toggle on
• Or say "Hey Siri, turn on VoiceOver"
• Or triple-click the side button if you've configured the Accessibility Shortcut under Settings > Accessibility > Accessibility Shortcut

Because VoiceOver alters how taps and swipes work, it can feel disorienting if you're not expecting it. The gesture changes are intentional and designed for users who navigate by audio rather than visual cues.

Android

Android's TTS features include TalkBack for full screen reading, Select to Speak for on-demand reading, and a system-level TTS engine that other apps can call.

Enabling Select to Speak

For most users, Select to Speak is the right starting point. It reads whatever you tap or select without changing your device's navigation.

• Go to Settings > Accessibility > Select to Speak
• Toggle it on
• A small overlay icon appears on screen. Tap it, then tap or drag across the text you want read aloud

On Samsung devices, the path may be Settings > Accessibility > Installed Apps > Select to Speak.

Enabling TalkBack

TalkBack is Android's equivalent of VoiceOver. It narrates every element and changes touch behavior to a select-then-activate model.

• Go to Settings > Accessibility > TalkBack and toggle on
• Or hold both volume buttons for 3 seconds on Android 9 and above to toggle TalkBack

Like VoiceOver on iOS, TalkBack changes how gestures work:

• Single tap selects and announces an item
• Double tap activates it
• Two-finger swipe scrolls the page
• One-finger swipe right or left moves to the next or previous element

Configuring the TTS Engine

Android lets you choose which TTS engine powers all speech output system-wide.

• Go to Settings > Accessibility > Text-to-Speech Output, or on Samsung devices, Settings > General Management > Language and Input > Text-to-Speech
• Select your preferred engine. Google's TTS engine comes pre-installed on most devices. Samsung offers its own alternative
• Tap the gear icon next to the engine to download additional language packs
• Use the "Speech rate" and "Pitch" sliders to customize how the voice sounds

Chromebook

ChromeOS keeps its TTS options in one place, making setup simpler than on most other platforms.

Enabling Select to Speak

• Go to Settings > Accessibility > Text-to-Speech
• Toggle on Select to Speak
• Click the Select to Speak icon in the system tray, then drag across any text on screen to hear it read

Enabling ChromeVox

ChromeVox is ChromeOS's full screen reader.

• Press Ctrl + Alt + Z to toggle ChromeVox on
• Or enable it under Settings > Accessibility > Text-to-Speech > ChromeVox

ChromeVox starts narrating immediately upon activation. It uses Google's TTS engine, same as Android, and supports the same language packs and voice options.

What Built-In Voices Get Right and Where They Fall Short

You've turned on TTS. Within about 30 seconds of listening, you'll notice the pattern.

Built-in voices handle short, simple sentences well. They pronounce common words correctly, pause at periods, and maintain a consistent speed. For reading a notification or a two-line text message, they're fine.

The cracks show up with longer content. Read a full article aloud using any built-in voice and listen for these signs:

• Flattened emphasis: Every sentence sounds the same. Important words get no extra stress. Questions don't rise in pitch the way a human voice would
• Awkward pacing around punctuation: Semicolons, colons, and parenthetical phrases confuse most engines. The voice either ignores them or inserts oddly long pauses
• Pronunciation drift: Technical terms, brand names, and foreign words get mangled. The voice commits to a pronunciation and repeats the same error every time it encounters the word
• Listener fatigue: After 2-3 minutes, the monotone quality becomes mentally draining. This is the main reason people turn TTS off shortly after turning it on

These aren't bugs. Built-in TTS engines are optimized for small file size, offline use, and universal compatibility. Sound quality is the concession.

AI Text to Speech Changes the Equation

If you turned on TTS hoping to listen to articles, proofread by ear, or produce voiceovers, and the built-in voice made you reconsider, the problem wasn't the feature. It was the engine.

AI voice platforms like Fish Audio use neural models trained on human speech. Instead of concatenating syllable fragments, these models generate audio from scratch, capturing the rhythm, emphasis, and tonal variation that make speech sound alive. The difference is obvious within the first sentence. Here's what Fish Audio's Text to Speech offers that device-level TTS doesn't:

• Stylistic controls: Industry-leading 64+ emotional and stylistic controls, covering nearly every expressive need from joy and sadness to anger and calm
• Natural prosody: The engine stresses words that matter, softens transitions, and varies pacing based on sentence structure. A question sounds like a question. A list sounds like a list. Built-in TTS reads everything with identical weight
• 13 languages with cross-language support: Switch between English, Mandarin, Spanish, Japanese, and more, even within the same paragraph, without pronunciation breaking down
• Browser-based workflow: No software to install. Go to fish.audio/text-to-speech, paste your text, choose a voice, and generate downloadable audio

Voice Cloning for Consistent Content

For creators who need the same voice across multiple projects, Fish Audio's Voice Cloning creates a custom model from as little as 10 seconds of reference audio. The model learns the speaker's tone, rhythm, and vocal texture, then applies those characteristics to any new text.

Practical applications include:

• YouTube and podcast production: Generate narration in a consistent voice without recording every script
• Multilingual content: A cloned voice retains its character when generating speech in different languages
• Brand voice consistency: Use the same voice across ads, tutorials, and customer communications without scheduling studio time for every update

API Access for Developers

Fish Audio's API exposes the full TTS and voice cloning engine for programmatic use. Response times are in the millisecond range with streaming support, which means real-time voice applications don't need to buffer.

Pricing and plan details are at fish.audio/plan. A free tier is available for testing.

Conclusion

Turning on text to speech takes under a minute on any platform. Win + Ctrl + Enter on Windows, Option + Esc on Mac, two-finger swipe on iPhone, Select to Speak on Android, Ctrl + Alt + Z on Chromebook. The feature is already on your device, waiting.

The harder question is whether you'll want to keep it on. Built-in voices work for quick reads and accessibility essentials, but they weren't designed for extended listening or content production. If the voice drives you to hit the off switch within two minutes, try Fish Audio's TTS before giving up on the concept entirely. The gap between a pre-installed engine and a modern AI voice is the difference between tolerating speech and actually preferring it to reading.