Skip to content
Lizely

Text To Speech

Read up to 20,000 characters aloud with a browser voice, adjustable rate and pitch, and explicit pause, resume, and stop controls.

Privacy: your files never leave your device. All processing happens locally in your browser.

How to use

  1. 1.Enter or paste the text you want the browser to read.
  2. 2.Choose an available voice and adjust rate or pitch if needed, then select Speak.
  3. 3.Use Pause, Resume, or Stop at any time; edit the text to start a fresh reading.

About Text To Speech

Text to Speech reads plain text with the speech engine already exposed by your browser and operating system. Paste or type up to 20,000 characters, choose an available voice when the platform reports one, set rate and pitch, and select Speak. Pause, Resume, and Stop remain explicit controls. The text stays in the current tab and is passed only to the local browser speech interface; Lizely does not upload or store it.

Long text is divided into bounded speech chunks of no more than 220 UTF-16 code units. The splitter prefers a sentence boundary, then a clause boundary, then a space, and finally a hard limit when one unbroken token is unusually long. Line endings and repeated horizontal spaces are normalized before playback, but words and punctuation are not rewritten. Chunking avoids sending one very large utterance to browser engines that may stop or behave inconsistently on long input. It does not provide linguistic sentence parsing or promise identical pauses across voices.

Rate ranges from 0.5× through 2.0× and pitch from 0.5 through 2.0. These are SpeechSynthesisUtterance control values, not measured words per minute, musical intervals, or accessibility certifications. Different engines interpret them differently. Voice names, languages, accents, quality, pronunciation, and availability come from the current device. A voice visible on one computer may be absent on another, and downloading extra operating-system voices is outside this tool.

Selecting Speak cancels any older queue created by this widget, creates one utterance for each current chunk, applies the chosen rate, pitch, and voice, and queues them in order. A generation number prevents callbacks from an older stopped reading from overwriting the status of a newer one. Editing text or a setting stops the active queue so the displayed summary cannot describe stale input. Unmounting the widget also cancels its queued speech.

The browser may require a user gesture before audio output, may expose no explicit voice list, or may reject a voice at runtime. The tool reports a playback error rather than pretending speech completed. Pause and resume are delegated to the same browser engine. They affect the global speech synthesis queue for the current page, so another script using that API in the same tab could interact with it. Lizely does not synthesize an audio file, expose a server voice, or guarantee offline availability.

Use this utility for listening to drafts, checking rhythm, hearing a short passage, supporting a reading workflow, or testing how a system voice pronounces text. It is not a screen reader replacement: it does not announce page structure, focus, controls, semantic roles, or live regions. It is also not a medical, diagnostic, language-learning, transcription, voice-cloning, or accessibility-conformance service. For essential assistive use, rely on a maintained screen reader and platform accessibility settings.

Playback output is not downloadable because the standardized browser speech interface exposes playback controls but no portable audio buffer. If you need a reusable recording, licensed commercial voice, fixed pronunciation dictionary, SSML workflow, or reproducible voice across machines, use a dedicated synthesis service and review its privacy and licensing terms. This tool deliberately keeps its scope narrow: local text, local voices, bounded utterances, and direct controls.

Methodology & sources

Reject blank text, null characters, and input above 20,000 UTF-16 code units. Normalize line endings and repeated horizontal spaces, then create chunks of at most 220 code units by preferring sentence punctuation, clause punctuation, whitespace, and finally the hard boundary. Validate rate and pitch inclusively from 0.5 through 2.0. Cancel the prior browser queue before each run; assign the selected locally reported voice to every utterance; guard start, end, and error callbacks with a generation number; and cancel queued speech on edits, Stop, failure, or unmount. No server request or audio-file encoding is performed.

Frequently asked questions

Is my text uploaded?
No. The text is passed to the speech synthesis interface in the current browser tab and is not sent to Lizely.
Why are the voices different on another device?
Voice names, languages, quality, and availability are supplied by the current browser and operating system, so they vary by platform.
Can I download an MP3 or WAV?
No. The browser Speech Synthesis API provides playback but does not expose a standardized audio buffer for download.
Is this a replacement for a screen reader?
No. It reads entered text only and does not provide the navigation, semantics, focus handling, or announcements of assistive screen-reader software.

Related tools

Device & Productivity guides

View all