Privacy‑First Text‑to‑Speech: What Stays On Your Device

Last updated: December 2025

People often paste sensitive text into text‑to‑speech tools: drafts, notes, letters, study material, even private messages. So it’s reasonable to ask: where does my text go?

Local Speech vs Cloud Speech

There are two common ways text‑to‑speech works:

Local (on-device / browser voices): speech is generated by your device’s speech engine.
Cloud TTS: your text is sent to a server, converted to audio, and the audio is returned.

For the deeper decision guide (privacy + voice quality + trust questions), see Local vs Cloud Text‑to‑Speech.

Tradeoff: Cloud TTS can provide very natural voices, but it requires sending text off-device. Read‑Aloud plays audio in your browser and does not export audio files, keeping control on your device.

How Read‑Aloud Handles Your Text

Read‑Aloud offers two types of voices with different privacy characteristics:

Browser voices: Your text stays completely on your device. Speech is generated by your browser/OS speech engine.
Premium neural voices: Text is sent to our TTS server for processing, then to Microsoft Edge TTS. Text is processed in real-time and not stored.

Read‑Aloud specifics

Browser voices come from your device. Different browsers/OSes expose different voice lists.
Premium voices require an internet connection and sound more natural.
On iPhone/iPad: if audio is silent or voices don't load, see Help.
No file downloads. Read‑Aloud focuses on in‑browser playback and does not export audio files.

Deep dives: Browser compatibility

What You Should Still Be Careful About

Your own device: if you’re on a shared computer, avoid pasting private content.
Browser extensions: some extensions can read page content. Use only extensions you trust.
Screen sharing: if you paste private text while sharing your screen, others can see it.

Privacy checklist (before you paste)

☐ I am not screen‑sharing (or the shared window does not include my text box)
☐ I am not on a shared/public computer
☐ I trust any browser extensions installed in this browser
☐ The text does not contain passwords, 2FA codes, or private identifiers
☐ If it’s sensitive, I will paste only the section I need (not the whole document)

Why downloadable audio files aren’t provided

Many system voices don’t provide a way for websites to capture their speech output as a file. The Web Speech API is designed to speak through your device speakers — it doesn’t hand a website an audio recording. That’s why Read‑Aloud focuses on private, in‑browser playback instead of file downloads.

Where to Learn More

Next: How to Use Read‑Aloud · All Guides