Text to Speech Reader

Proofreading by eye misses what proofreading by ear catches, because your eyes autocorrect a skipped word and your ears don't. Text to Speech hands the reading over to whatever voice is already installed on your device, using the Web Speech Synthesis API that ships in every modern browser rather than a hosted narration service.

The word being spoken is highlighted in a live preview as it plays, which is the detail that turns this from a novelty into something worth opening twice. Nothing is uploaded to produce it. The audio is generated by the operating system's own speech engine, running in the tab you already have open.

Getting Your First Sentence Read Back

1. Press Play — the box already holds “This is what your browser's built-in voice sounds like…”. Press Load Sample first if you would rather hear a different script.
2. Pick a different entry from the Voice dropdown mid-sentence and press Play again. The change only applies to the next utterance, not the one already speaking.
3. Drag the Rate slider past 1.5 for a fast skim of a long paragraph, or under 0.8 to catch a line you keep mis-reading.
4. Pause and Resume use the browser's own pause state rather than stopping and restarting the utterance, so resuming picks up mid-word instead of from the beginning.

What Happens When a Voice Isn't Installed

• A browser with no speech engine at all shows Speech synthesis isn't available in this browser. Try a recent Chrome, Edge, Safari or Firefox build. instead of a dead Play button.
• Voice lists load asynchronously in Chrome-based browsers, so the dropdown briefly reads Loading voices… before the real list of installed voices replaces it — pressing Play during that window still works, it just uses whichever voice loads first.
• Pasting past 3,000 characters raises Only the first 3,000 characters will be read — long-form narration needs a dedicated app, not a browser tab. The counter under the box updates on every keystroke so the cut-off is never a surprise.

The Limits of a Browser's Own Voice

Word-level highlighting depends on the browser reporting a boundary event with a character length attached. Not every engine sends one. When it's missing, the highlight falls back to measuring the next run of non-space characters instead, which is accurate for ordinary prose and occasionally off by a character on hyphenated compounds.

• Voice quality and the number of installed voices depend entirely on the operating system, not on this page — the same text can sound noticeably different read on Windows, macOS and a phone.
• Rate runs from 0.5× to 2×, pitch from 0 to 2, and volume from 0 to 1, matching the ranges the underlying SpeechSynthesisUtterance API actually accepts. Values outside them are simply not offered on the sliders.
• Switching browser tabs mid-sentence can pause playback on some mobile browsers regardless of what the Pause button shows, since the operating system, not this page, owns that decision.