Skip to main content

text-to-speech

A text-to-speech (TTS) tool synthesises spoken audio from written text using a voice engine β€” letting you listen instead of read, generate voiceovers, proof-listen drafts, or build accessibility experiences. The ZTools Text to Speech runs entirely in the browser using the Web Speech API's SpeechSynthesis interface, exposing every system-installed voice (typically 20–80+ voices spanning 30+ languages on modern OSes), with adjustable rate, pitch, and volume. No audio leaves your device, no API quota, no signup β€” the synthesiser is the same one your operating system already ships.

Use cases​

Proofreading by ear​

Reading your own writing silently misses awkward phrasing the brain auto-corrects. Listening surfaces typos, run-on sentences, and rhythm problems that visual proofreading skips. Faster than re-reading carefully.

Accessibility for low-vision users​

Quick TTS for documents, emails, articles when full screen-reader software is overkill. Paste, listen, move on.

Language learning pronunciation​

Hear how a phrase sounds in the target language. Switch voices to compare regional accents (en-US vs en-GB, es-ES vs es-MX). Slower rate helps with new vocabulary.

Voiceover prototypes​

Quick-and-dirty narration for video drafts, presentation timing tests, or e-learning prototypes before recording with a real voice actor.

Multitasking​

Listen to a long article while cooking or commuting. Faster onboarding to long content than reading from start to finish at a desk.

How it works​

  1. Paste or type text β€” Up to ~32k characters per utterance is safe across browsers; long inputs are auto-chunked at sentence boundaries.
  2. Pick a voice β€” Dropdown lists every voice your OS exposes via SpeechSynthesis.getVoices() β€” language, gender, and engine (Google, Microsoft, Apple) shown.
  3. Adjust rate & pitch β€” Rate 0.1–10 (default 1.0); pitch 0–2 (default 1.0). Volume 0–1.
  4. Press Speak β€” SpeechSynthesisUtterance fires; pause/resume/stop controls available mid-speech.
  5. Optionally record β€” On supported browsers, capture the synthesised audio via MediaRecorder for download as .webm/.wav.

Examples​

Input: "The quick brown fox jumps over the lazy dog." Voice: en-GB, rate 0.9.

Output: Slow, clearly enunciated British English audio β€” useful for dictation practice.

Input: Long-form blog draft (~3000 words). Voice: en-US, rate 1.2.

Output: Auto-chunked into ~50 utterances at sentence boundaries; total runtime ~12 minutes.

Input: "γ“γ‚“γ«γ‘γ―γ€δ»Šζ—₯はいい倩気ですね。" Voice: ja-JP.

Output: Native Japanese pronunciation; useful for learners who can read kana but want to hear it spoken.

Frequently asked questions​

Is this the same as ElevenLabs / Google Cloud TTS?

No β€” those are paid neural-voice APIs producing studio-quality audio. ZTools uses your browser/OS's built-in synthesiser, which is free and instant but sounds more robotic. Trade-off: quality vs cost.

Why are some voices missing?

Available voices come from your OS, not from us. Windows ships fewer voices than macOS by default; many languages need an OS-level language pack install. Chrome on Linux often lists fewer voices than Chrome on Windows.

Does it work offline?

OS-installed voices work offline; cloud voices (e.g. Chrome's "Google" voices) need a network connection because the synthesis happens server-side.

Can I download the audio?

Yes on browsers that allow capturing the audio output stream β€” typically via MediaRecorder. Some browsers block this for cloud-synthesised voices for licensing reasons.

Why does it stop after ~250 characters?

A known Chrome bug on long utterances β€” workaround is to chunk at sentence boundaries (the tool does this automatically).

Can I add SSML (pauses, emphasis)?

The Web Speech API supports a small subset of SSML on some browsers but it is inconsistent. Use commas, periods, and ellipses for natural pauses instead.

Tips​

  • Pick the OS-bundled voices for offline use; they are typically named "Microsoft <Name>", "Apple <Name>" β€” these don't need network.
  • Lower the rate (0.85–0.9) for proofreading; you'll catch more issues than at default speed.
  • For long documents, split at chapter breaks and queue utterances β€” avoids the Chrome long-input bug.
  • Test in multiple browsers β€” voice availability differs significantly between Chrome, Edge, Safari, and Firefox.
  • For production voiceover, use this for timing/pacing prototypes only and re-record with a paid neural voice service.

Try it now​

The full text-to-speech runs in your browser at https://ztools.zaions.com/text-to-speech β€” no signup, no upload, no data leaves your device.

Open the tool β†—

Last updated: 2026-05-06 Β· Author: Ahsan Mahmood Β· Edit this page on GitHub