Privacy Policy — Select to Speak

Last updated: August 23, 2026

Select to Speak ("the extension") reads text you select in your browser aloud. This policy explains exactly what the extension does and does not do with data. In short: the extension has no servers, no accounts, and no analytics, and it does not send your data to its developer or to any third party for tracking or advertising.

What the extension accesses

• Selected text. When you choose "Speak selection", press the keyboard shortcut, or use the playback controls, the extension reads the text you have highlighted so it can be spoken. This text is used only to produce speech and is not stored or logged by the extension.
• Your settings. Your chosen voice, speed, pitch, volume, and theme are saved using the browser's chrome.storage.sync feature so they persist and follow your browser profile. These values are stored by your browser (and, if you have browser sync enabled, synced through your browser account); they are never sent to the developer.

How your selected text is processed

Speech is produced by a text-to-speech engine. Where your selected text goes depends on which voice you use:

• On-device (neural) voices and offline system voices synthesize speech entirely on your computer. Your selected text never leaves your device.
• Online system voices (for example, voices labeled "Online" or provided by your operating system or browser that require an internet connection) are synthesized by that voice provider's servers. When you use such a voice, your browser sends the selected text to that provider in order to speak it. This transmission is performed by your browser using the operating system's or browser's own voices, and is governed by that provider's privacy policy. To keep everything local, choose an on-device neural voice or an offline system voice.

Network requests the extension makes

The extension itself makes exactly one kind of network request: when you choose to download an on-device neural voice, it downloads that voice's model file from the public Hugging Face model repository. Only the model file is downloaded — no selected text, settings, or personal data are sent with these requests. Once a voice is downloaded, it works offline.

Data the extension does NOT collect

• No personally identifiable information.
• No browsing history, page content (beyond the text you explicitly select to be spoken), or form data.
• No credentials, financial, health, or location data.
• No analytics, telemetry, tracking, or advertising identifiers.
• Nothing is sold or transferred to third parties.

Permissions

The extension requests only the permissions needed to function:

• contextMenus — to add the right-click menu items.
• tts — to speak text using the browser's text-to-speech engine.
• storage — to save your voice and playback settings.
• offscreen — to play audio for on-device neural voices.
• scripting and activeTab — to read the current text selection when you use the keyboard shortcut (granted only for the active tab, only when you invoke the command).
• host access to huggingface.co — solely to download neural voice model files you choose.

Children's privacy

The extension does not collect any personal data from anyone, including children.

Changes to this policy

If this policy changes, the updated version will be posted at this URL with a new "Last updated" date.

Contact

Questions about this policy: open an issue on GitHub.