Speech synthesis voices
Installed text-to-speech voices, which reveal OS, version and installed language packs.
In one line
Can list the speech voices on your device, which hint at your OS and languages
Installed text-to-speech voices, which reveal OS, version and installed language packs.
macOS lists 100+ voices, Windows 10 to 30, Android per Google TTS pack. Exact list shows OS version and which languages the person installed, a strong stable fingerprint.
What it can and can't do
What the user does: Nothing. It happens when the page loads or speaks text
It can
- List installed voices, from about 10 on Windows to 100 or more on macOS
- Guess your OS version and language packs from the list
- Offer a speak text feature in a voice you pick
- Use the list as a steady tracking clue
It cannot
- Listen to you or record sound
- See your speech or text you wrote
- Know which voice you like best
- Get the same list on every browser or device
Browser support
| Browser | Status | First version | Released | Note |
|---|---|---|---|---|
| Chrome | Supported | 33 | 2014-02-20 | None |
| Edge | Supported | 14 | 2016-08-02 | None |
| Firefox | Supported | 49 | 2016-09-20 | None |
| Safari | Supported | 7 | 2013-10-22 | None |
| Safari on iPhone | Supported | 7 | 2013-09-18 | None |
| Chrome on Android | Supported | 33 | 2014-02-26 | None |
| Samsung Internet | Supported | 3.0 | 2015-04-10 | None |
How a site asks for it
| API | speechSynthesis.getVoices() |
| Browser compat keys | api.SpeechSynthesis.getVoices |
| How access is granted | none |
| What the user sees | None |
| Needs a secure context (HTTPS) | No |
| Needs a user gesture | No |
| How long it lasts | read fresh on every page load; nothing stored |
| How to take it back | none needed (no grant) |
| Inside an iframe | Readable from cross-origin iframes too (same values) |
Typical flow
- Page script reads the value
- No prompt is shown
Data you get
Shape: object
How often: static per session
| Field | Type | Unit | Range | Precision |
|---|---|---|---|---|
name | string | None | e.g. Samantha, Microsoft David, Google US English | Not listed |
lang | string | BCP 47 | e.g. en-US | Not listed |
localService | boolean | None | true or false | Not listed |
voiceURI | string | None | OS-specific | Not listed |
default | boolean | None | Not listed | Not listed |
Technical limits
- List may load asynchronously (voiceschanged)
- Safari 26 blocks known fingerprinters
The story
What it is
The list of text to speech voices installed on your device, which reveals system and languages.
For you
You never see the check. If a site reads text aloud, it can use a voice you have installed. The same list can show which languages you added and what system you run.
For makers
Reading aloud helps accessibility and language learning. The voice list changes by system, version and language pack, so it is also a strong fingerprint. Trackers read it for that reason.
Where you see it
| Type | Example | Apps |
|---|---|---|
| accessibility | A news site offers a Listen button that reads the story aloud in the voice you have installed. | None named |
| education | A language learning site picks a native speaker style voice for your language to read sentences aloud. | None named |
| games | A browser story game reads dialogue out loud with the best voice available on the device. | None named |
| kids | A children's reading site speaks each word aloud so early readers can follow. | None named |
| productivity | A web editor reads your draft back so you can hear mistakes. | None named |
| security | A fingerprint vendor reads the voice list as a stable ID piece. | FingerprintJS |
| analytics | A traffic tool estimates language packs installed to judge international audiences. | None named |
Risks
- Strong stable fingerprint
- Reveals languages you added
What the platform says
- MDN: Web Speech API. Docs describe how pages can speak text aloud using voices on the device.
Good uses and abuse
Legitimate uses
- Pick voice for reading aloud
Known abuse
- OS and language-pack fingerprint
Scores
| Score | Value | Meaning |
|---|---|---|
| Sensitivity | 1 of 5 | How revealing the data is |
| Utility | 3 of 5 | How useful it is to honest apps |
| Friction | 0 of 5 | How much the user has to do |
| Fingerprint bits | 5 | Rough identifying entropy, when known |
History
- 2025-09, Safari 26: Voice list restricted for known fingerprinting scripts (Advanced Fingerprinting Protection)
Related capabilities
Related
- Installed fonts via measurement. Can tell which fonts you have installed by measuring text, with no prompt
Sources
How well checked: Inferred from related behavior.