Text to Speech in Your Browser: How It Works (No Upload, No Account)

Every modern browser can read text aloud — using the same voices your device uses for accessibility — without an account, an upload, or (for local voices) an internet connection. It’s built on the Web Speech API, it’s free, and it’s quietly private, if you pick a local voice. This guide explains how it works, which voices stay on your device versus which ones don’t, why the voice list looks different on every machine, what it’s genuinely good for, and its one real limitation — using the Text to Speech tool.

The browser already has a voice

Behind the feature is the Web Speech API’s speech synthesis, built into Chrome, Edge, Safari and Firefox. It taps the text-to-speech voices installed in your operating system — the same ones that power screen readers and “speak selection.” So reading text aloud needs no plugin and no third-party service: you hand the browser some text and it speaks it, using a voice from a list the browser gets from your OS.

A good TTS tool just wraps that engine with controls: pick a voice and language, set the rate, pitch and volume, and play, pause or stop.

Which voices are private (and which aren’t)

This is the part most guides get wrong. Whether your text stays on your device depends on the specific voice you choose:

• Local voices (localService = true) — installed in your OS, they synthesise audio entirely on your machine. Nothing is uploaded. These are the private ones.
• Network voices (localService = false) — most notably Chrome’s built-in “Google” voices, which generate the audio on Google’s servers. Choosing one sends your text over the network.

So in-browser TTS can be fully private — just prefer a voice marked local. That’s the opposite of many “free TTS” websites, which stream your text to a cloud engine no matter what and gate it behind sign-up and quotas. Local voices keep your words on your machine and work offline.

Why the voice list differs everywhere

The voices come from your OS and browser, not the website, so the dropdown looks different on every device:

Platform	Typical built-in voices
Windows	David, Zira, Mark + downloadable “natural” voices
macOS	Alex, Samantha, Daniel + dozens of language voices
iOS / iPadOS	Siri-quality voices, many languages
Android	Google TTS voices (varies by device)
Chrome (any OS)	Adds network “Google” voices to the OS list

The upside: you can install more language voices in your operating system’s settings, including high-quality “natural/enhanced” ones, and they appear in the tool automatically. If you want a specific accent or language, add it at the OS level.

Getting good results

• Proofreading — listen to a draft at normal speed; your ear catches missing words, doubled words and clumsy sentences the eye skips right over. It’s the fastest proofreading trick there is.
• Accessibility — for anyone who finds listening easier than reading, or wants to rest their eyes.
• Language learning — pick a voice in the target language to hear pronunciation of words and phrases.
• Skimming vs. studying — raise the rate to 1.3–1.6× to power through an article, or slow it to follow along.
• Punctuate for pauses — commas and periods shape the phrasing, so well-punctuated text reads more naturally.

Downloading audio: two different voices

The browser’s OS voices are built for playback, not capture — the Web Speech API never hands their audio to JavaScript, so those natural voices genuinely can’t be saved to a file. But that isn’t the end of the story. To offer a real download without giving up privacy, the Text to Speech tool includes a Download WAV button powered by a small offline speech engine (eSpeak-NG/meSpeak) compiled to run in your browser. Unlike the Web Speech API, that engine produces the audio samples in JavaScript, so it can build a WAV file and hand it to you — all on your device, nothing uploaded.

The trade-off is voice quality: the downloadable offline voice sounds more robotic than the natural OS voices you hear on Play. So the tool uses the best of both — natural voices for listening, the offline engine for a genuine downloadable file. The engine (about 1 MB) loads only the first time you click download, so it never slows down normal use.

Listen to any text, privately

The Text to Speech tool reads your text aloud with your device’s voices — choose the voice and language, tune the speed and pitch, and play, pause or stop — with nothing uploaded by the tool and no account needed. Pick a local voice, paste a draft, and hear it: your ear will catch what your eyes glossed over.